White matter tract involvement in brain tumors: a diffusion tensor imaging analysis

Available online at www.sciencedirect.com

Surgical Neurology 72 (2009) 464 – 469 www.surgicalneurology-online.com

Imaging

White matter tract involvement in brain tumors: a diffusion tensor imaging analysis☆ Pao Sheng Yen, MD a,⁎, Beng Tiong Teo, MD b , Cheng Hui Chiu, MD a , Shang Chi Chen, MD b , Tsung Lang Chiu, MD b , Chain Fa Su, MD b Departments of aMedical Imaging and bNeurosurgery, Buddhist Tzu Chi General Hospital and Tzu Chi University, Hualien, Taiwan Received 15 December 2008; accepted 4 May 2009

Abstract

Background: Characterization of WM alteration using MR imaging is important in the pre- and intraoperative assessment of brain tumors. This study characterizes the extent and severity of WM tract alterations near brain tumors using DTI in an effort to determine preoperative viability or resectability of the adjacent WM tracts. Fractional anisotropy is an important DTI-derived metric of MR imaging. Methods: Twenty-one patients underwent MR DTI. Eighty-six WM tracts composed of 43 WM lesions paired with 43 contralateral WM hemispheric controls were categorized using FA. Neuroradiologists categorized the WM tracts as edematous, displaced, disrupted, or infiltrated with tumor using directionally encoded color maps. A mixed model analysis was used to compare FA. Results: Of the lesioned tracts, 5 were scored as edema, 9 as infiltration, 18 as displacement, and 11 as disruption. A significant ΔFA% was found between the lesioned and contralateral hemispheres only in WM disruption (P = .0056). Both edema FA and disruption FA are significantly less than displacement FA (P b .05). The FA change (ΔFA% = [FAlesion − FAnormal]/FAnormal × 100%) on the lesioned side was calculated. A ΔFA% less than −30% is likely to be associated with WM disruption. A positive ΔFA% is likely to be associated with edema or displacement, and a ΔFA% between 0% and −30% is likely to be associated with WM displacement or infiltration. Conclusions: Quantitative analysis of DTI data may provide insight as to whether WM tracts are salvageable preoperatively. © 2009 Elsevier Inc. All rights reserved.

Keywords:

White matter neoplasm; Diffusion tensor imaging

1. Introduction Magnetic resonance DTI provides information useful to the clinician concerning WM integrity and organization in the brain [1-4,6,8,10,15,18-22,24,26] by identifying WM Abbreviations: DTI, diffusion tensor imaging; EPI, echo planar imaging; FA, fractional anisotropy; FLAIR, fluid-attenuated inversion recovery; ΔFA%, FA difference; IRB, institutional review board; MD, mean diffusivity; MR, magnetic resonance; PNET, primitive neuroectodermal tumor; ROI, region of interest; WHO, World Health Organization; WM, white matter. ☆ This work was supported by grant from the Hualien Tzu Chi General Hospital medical research fund (TCRD 94-17). ⁎ Corresponding author. Tel.: +886 930332007; fax: +886 38571287. E-mail address: [email protected] (P.S. Yen). 0090-3019/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.surneu.2009.05.008

tract pathology with greater sensitivity [2,18,19,26] and characterizing WM injury and tract reorganization after treatment [3,7,10,24]. Diffusion tensor imaging has also been used to differentiate high- from low-grade brain tumors [4,20-22] and to determine tumor histology before biopsy [1,8]. In addition, DTI can be used to plan the radiotherapy target volume to include regions of tumor cell invasion [6], as well as to better delineate surgical tumor margins [2,15]. Clear delineation of tumor margins is crucial because surgical resection of brain tumors requires a delicate balance between maximal resection of tumor tissue and minimal injury to surrounding normal WM. White matter tract involvement adjacent to brain tumor can be

P.S. Yen et al. / Surgical Neurology 72 (2009) 464–469

465

classified as edema, infiltration, displacement, or disruption [25]. There has been an effort to distinguish between these tissue types using T2-weighted images or DTI [1,3,4,15,18,19,21,25]. Yet preoperative determination of surgical margins for resection remains a major challenge in neurosurgical oncology. Fractional anisotropy and MD are both important DTIderived metrics, and accurate quantification of both FA and MD could potentially improve the characterization of WM lesions by DTI [11,13,14]. We use FA to characterize WM tracts adjacent to brain tumor in an effort to determine preoperatively whether a specific tract is viable or should be resected during surgery. We test the hypothesis that DTI can preoperatively differentiate among edema, infiltration, displacement, and disruption of WM tracts by tumor. 2. Materials and methods A total of 21 patients with brain lesions were evaluated between April 2004 and May 2006. Among these 21 patients, a total of 43 patients had potentially abnormal WM tracts identified adjacent to a lesion, and these WM tracts were compared to WM tracts in the contralateral hemisphere, which were assumed to be normal. Because all patients underwent a routine preoperative MR examination, which included T1- and T2-weighted sequences, IRB permission was not required to add a DTI sequence to the patient examination.

Fig. 1. Diffusion tensor imaging directionally encoded color map in a 34year-old patient who presented with generalized seizure. The T2-weighted image (A) revealed a left frontal mass with marked perifocal hyperintensity. The DTI directionally encoded color map (B) showed markedly decreased hue of the anterior limb of the internal capsule (AL) (solid arrow), although an infiltrative or destructive process cannot be differentiated. The hollow arrow may represent displaced fiber and the center of the lesion may represent disruption.

Fig. 2. Average FA of individual WM tracts. After drawing a ROI in the anterior limb of the internal capsule on both hemispheres of the directionally encoded color map (A), ROIs were transferred to the FA map (B). The average FA of individual WM tracts was then measured. ΔFA% was calculated from the averaged FA values of the lesion and normal hemisphere using Eq. (1). For this particular case, ΔFA% of the anterior limb of the internal capsule was [(0.3512 − 0.5175)/0.5175 × 100%] = −32.13%.

Diffusion tensor imaging was performed using a singleshot spin-echo EPI sequence with standard parameters (TR/ TE = 8000/100, FOV = 28 cm, acquisition matrix = 128 by 128). Section thickness was 4.4 mm with 0-mm spacing because the postprocessing software requires images of this format, and images were acquired to cover the entire brain (approximately 28 images). Diffusion-sensitizing gradient encoding was applied in 25 directions, with a diffusionweighting factor b value of 1000 s/mm2, and one image was acquired without diffusion-sensitizing gradients (ie, b = 0 s/ mm2). Total imaging time was approximately 8 minutes. The DTI images, including FA and directionally encoded color maps, were reconstructed using software provided by the VOLUME-ONE developer group, which is freely available on the internet (http://www.volume-one.org). We identified individual WM tracts using directionally encoded color maps [11,13]. Hues in the color map represent tensor direction, whereas grayscale intensity represents the integrity of WM tracts. The location of individual WM tracts was identified by reference to a pictorial review based on cadaveric dissections [5]. White matter tract involvement was categorized as edema, infiltration, displacement, or disruption [16]. According to the method of Witwer et al [25], WM tracts were characterized as displaced “if they maintained normal anisotropy relative to the corresponding tract in the contralateral hemisphere but were situated in an abnormal location or with an abnormal orientation on color-coded orientation maps; edematous if they maintained normal anisotropy and orientation but demonstrated high signal intensity on T2-weighted MR images; infiltrated if they showed reduced anisotropy but remained identifiable on

466

P.S. Yen et al. / Surgical Neurology 72 (2009) 464–469

orientation maps; and disrupted if anisotropy was markedly reduced such that the tract could not be identified on orientation maps.” Categorization was based on the 3-dimensional appearance of the tract (Fig. 1) and on pathology of the resected tumor tissue. Meningioma and neuroma were defined as noninfiltrative. Fractional anisotropy was then measured in every WM tract adjacent to tumor, as well as in the corresponding WM tract in the contralateral hemisphere. A ROI was drawn manually on the 2-dimensional image (Fig. 2) using VOLUME-ONE software.

The ΔFA% on the lesioned side was defined as the percentage of FA decrement adjacent to the tumor, compared to the contralateral “normal” hemisphere: DFAk =

FAlesion
FAnormal  100k FAnormal

ð1Þ

Use of FAnormal as the denominator eliminates discrepancies between different individuals or different WM tracts that might arise from differences in the way the DTI examination was performed.

Table 1 Fractional anisotropy values of WM adjacent to a lesion and in normal WM in the contralateral hemisphere Case

7 8 17 1 2 3 4

9 10 12 13 15 16 20 4 12 13 14 17 18 20 5 6 10 11 13 18 19 21

Age

67 67 55 55 64 44 51 80 80 58 58 58 58 37 37 31 31 37 18 18 53 53 50 58 31 37 37 29 64 64 69 50 44 67 37 13 13 37 37 69 23 22 22

Sex

M M F F M F F F F F F F F F M M M F F F M M M F M F F F M M M M F M M M M F F M F F F

Diagnosis

Metastasis Metastasis Gliomatosis cerebri Meningioma Meningioma Meningioma Meningioma

Metastasis Pontine glioma; arachnoid cyst Glioblastoma multiforme Glioblastoma multiforme Acoustic neuroma Acoustic neuroma PNET Meningioma Glioblastoma multiforme Glioblastoma multiforme Trigeminal neuroma Gliomatosis cerebri Gliomatosis cerebri Metastasis Metastasis Pontine glioma; arachnoid cyst Pilocytic astrocytoma Glioblastoma multiforme Oligodendroglioma Ependymoma Gliosis

WM tracts involved

Characteristics of WM involvement

FA side Lesion

Normal

ΔFA%

L ILF R CST L IOF L OR L Unc L CST LF L Unc L CST R OR R CST R SLF R SOF L OR L Unc L CST L OR L IOF L CP L CST R CP R CST R CST R IOF L IOF R SLF R Cin L CP L IOF L CST R OR R CST R CST R ILF CTT L IOF L CST R CST R CST R IOF R IOF R ILF R IOF

Edema Edema Edema Edema Edema Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Displacement Infiltration Infiltration Infiltration Infiltration Infiltration Infiltration Infiltration Infiltration Infiltration Disruption Disruption Disruption Disruption Disruption Disruption Disruption Disruption Disruption Disruption Disruption

0.3122 0.3892 0.3666 0.3740 0.2017 0.4087 0.3762 0.4631 0.4745 0.5489 0.4754 0.4018 0.5720 0.3385 0.2474 0.5633 0.4018 0.3219 0.4661 0.3788 0.4373 0.4323 0.4649 0.3768 0.2424 0.3961 0.2803 0.3682 0.2535 0.4995 0.3136 0.3975 0.2819 0.2188 0.3255 0.0000 0.3196 0.2117 0.3273 0.2705 0.2604 0.1889 0.2383

0.2647 0.3870 0.3592 0.3532 0.2520 0.4166 0.3449 0.4752 0.4240 0.6041 0.6115 0.3790 0.4453 0.3910 0.3082 0.5434 0.4332 0.3371 0.4962 0.5051 0.4854 0.5150 0.6363 0.3375 0.4194 0.4776 0.3043 0.5546 0.3100 0.4682 0.3773 0.4806 0.3716 0.3776 0.5141 0.2469 0.3895 0.2941 0.4119 0.4197 0.4046 0.4128 0.3037

17.94 0.57 2.06 5.89 −19.96 −1.90 9.08 −2.55 11.91 −9.14 −22.26 6.02 28.45 −13.43 −19.73 3.66 −7.25 −4.51 −6.07 −25.00 −9.91 −16.06 −26.94 11.64 −42.20 −17.06 −7.89 −33.61 −18.23 6.69 −16.88 −17.29 −24.14 −42.06 −36.69 −100.00 −17.95 −28.02 −20.54 −35.55 −35.64 −54.24 −21.53

R indicates right; L, left; CST, corticospinal tract; F, frontal; Unc, uncinate fasciculus ; IOF, inferior occipitofrontal fasciculus; OR, optic radiation; SLF, superior longitudinal fasciculus; ILF, inferior longitudinal fasciculus; SOF, superior occipitofrontal fasciculus; CTT, central tegmental tract; Cin, cingulum; CP, cerebellar peduncle.

P.S. Yen et al. / Surgical Neurology 72 (2009) 464–469 Table 2 Characteristics of WM involvement in 43 WM tracts studied Type of involvement

Brain lesion type

No. of WM tracts studied

Edema

Metastasis Gliomatosis cerebri Glioblastoma multiforme Gliomatosis cerebri Extra-axial tumors Oligodendroglioma PNET Extra-axial tumors Glioblastoma multiforme PNET Pontine glioma Metastasis Glioblastoma multiforme Metastasis Pilocytc astrocytoma Gliosis Oligodendroglioma Ependymoma Pontine glioma

4 1 3 2 2 1 1 12 3 1 1 1 2 2 2 2 1 1 1

Infiltration

Displacement

Disruption

2.1. Statistical analysis

467

the tracts qualitatively, there were 5 WM tracts with edema, 9 with infiltration, 18 with displacement, and 11 with disruption (Table 2). The FA values of the affected and the contralateral WM tracts are listed in Table 1. We verified that we were interrogating the same WM tract on the affected as the unaffected side by using the RGB maps because different WM tracts will have different color (direction). We compared the FA value of WM pairs using a mixed model approach to compensate for repeated measurements (Table 3). A significant ΔFA% was noted only for WM disruption. There was no significant ΔFA% between WM in lesioned and normal hemispheres for edema, displacement, or infiltration. The ΔFA% associated with each type of WM alteration is shown in Table 4. A ΔFA% less than −30% is likely to be associated with WM disruption (Table 4). In contrast, a positive ΔFA% is likely to be associated with edema or displacement. Finally, a ΔFA% between 0% and −30% is likely to be associated with WM displacement or infiltration (Table 4). 4. Discussion

The FA from lesioned and normal hemispheres was expressed as mean ± SD. A mixed model analysis was used, corrected for repeated measures with the Bonferroni method. Statistical calculations were performed using SAS version 9.0 (SAS Institute, Cary, NC), with a significance level of .05. 3. Results The WM tracts of 21 patients with WHO-graded brain tumors were evaluated, including 4 patients with meningioma, 5 with brain metastases, 2 with low-grade glioma, 2 with glioblastoma multiforme, 3 with neuroma, 1 with high-grade glioma, 1 with oligodendroglioma, 1 with ependymoma, 1 with PNET, and 1 with gliosis who was suspected to have recurrent astrocytoma. A total of 43 tracts near a lesion were compared to the contralateral (normal) hemisphere; therefore, 86 tracts were evaluated. By category, there were 15 corticospinal tracts involved, 13 WM tracts in the frontal lobe, 6 in the temporal lobe, 5 in the optic radiation, and 4 bulbar tracts (Table 1). Grouping

We characterized the extent and severity of WM tract alterations near brain tumors using DTI to test the hypothesis that DTI can differentiate among edema, infiltration, displacement, and disruption of WM tracts by tumor. Diffusion tensor imaging was performed in 21 patients with a variety of tumors. Using directionally encoded color maps, WM tract involvement was categorized as edema, infiltration, displacement, or disruption according to the method of Witwer et al [25]. We found that there was a significant ΔFA% between lesioned and contralateral hemispheres when there was WM disruption (P b .05). This suggests that WM integrity is unaffected by edema, infiltration, or displacement, although FA is significantly decreased by disruption. Our results suggest that disrupted WM tracts are at least partially destroyed. In addition, our findings may provide a method whereby salvageable WM tracts can be differentiated from WM tracts damaged or invaded by tumor before surgery. Inadvertent injury to WM tracts adjacent to a cortical area can be catastrophic during a neurosurgical procedure [17].

Table 3 Comparison of FA value on the lesioned side to the contralateral side

Table 4 Characteristics of WM involvement in calculated ΔFA%, analyzed by strata

Type of WM involvement

Normal

Lesion

P

Type of WM involvement

ΔFA% strata

Edema Displacement a,b Infiltration Disruption

0.32 ± 0.06 0.46 ± 0.10 0.41 ± 0.09 0.38 ± 0.07

0.33 ± 0.07 0.43 ± 0.08 0.35 ± 0.08 0.24 ± 0.09

.8667 .1759 .1470 .0056 ⁎

Less than −30

−30 to 0

N0

Edema Displacement Infiltration Disruption

0 0 2 6

1 13 5 5

4 (36.36) 5 (45.46) 2 (18.18) 0 (0)

Edema FA b displacement FA. Displacement FA N disruption FA. ⁎ P b.05. a

b

(0) (0) (25.00) (75.00)

P

(4.17) (54.17) (20.83) (20.83)

Numbers in parentheses represent percentage within the category. ⁎ P b.05 (Fisher exact test).

b.0001⁎

468

P.S. Yen et al. / Surgical Neurology 72 (2009) 464–469

Therefore, a detailed understanding of WM tracts near a brain tumor before surgery is crucial. Intraoperative stimulation mapping has been used to mitigate injury to both cortical and subcortical areas [17]. Until recently, this intraoperative technique was the only technique able to reliably delineate the descending subcortical motor, sensory, and language tracts. However, DTI is now becoming widely used to depict WM tracts intraoperatively [12] and our results suggest that we may be able to extend this method to differentiate salvageable WM tracts preoperatively. Many studies have attempted to delineate WM tract anatomy with DTI [1-4,6-8,10,11,15,18-22,24-26]. Directionally encoded color mapping is the simplest method because the hues represent tensor direction and the intensity represents WM integrity. Color maps thus summarize DTI data in an easily interpretable format. Using such directionally encoded color maps, 4 patterns of WM alteration have been described adjacent to tumor including edema, infiltration, displacement, and disruption [25]. However, early DTI studies were unable to discern between vasogenic edema and tumor infiltration of WM [1,21,25]. Tumor cell invasion into surrounding normal brain is a major clinical problem. A recent study [5], which described WM anatomy based on cadaveric dissection, found that there were 4 patterns of altered WM associated with brain tumors: normal FA but abnormal WM tract location or direction; decreased FA but normal WM tract location or direction; decreased FA and abnormal WM tract location or direction; and near-isotropic diffusion such that WM tracts could not be identified. It is likely that these 4 patterns correspond to specific patient outcomes, although this has yet to be proven. Diffusion tensor imaging alone, however, cannot give insight into these distinctions because anatomically intact WM fibers may be present in abnormalappearing areas of the brain [25]. Some combination of the various types of MR imaging, in the future, may provide insight into WM integrity in brain tumor patients. For example, specific WM tracts could be identified by tractography, the 3-dimensional representation of WM tracts based on DTI data. An independent method such as FLAIR imaging could then be used to highlight vasogenic edema [23]. Combining DTI with FLAIR may make it possible to differentiate between normal WM (even if surrounded by vasogenic edema) and disrupted WM that is invaded by tumor cells. Multiparametric imaging may enable clinicians to differentiate between neoplasm, edema, and healthy WM [16,23]. In-depth knowledge of WM integrity adjacent to a tumor could potentially alter surgical approach, the extent of attempted resection, or patient prognosis before surgery. There are several limitations of our study. Slice thickness was 4.4 mm for the DTI acquisition, which is rather thick; acquisition time was 8 minutes, so image quality could be degraded by patient motion. Regions of interest were drawn manually, which may be less easily reproduced than an automated method. Average FA was calculated based on

RGB maps rather than on an independent method such as T2-weighted imaging; and the actual size of WM tracts could not be determined by the fiber-tracking technique [9]. Using MR imaging alone, it is very difficult to differentiate between infiltration and disruption of WM tracts; therefore, we calculated the FA value to verify the results. We could not be sure that there was only infiltration or only displacement of the adjacent WM tracts in our cases. Only autopsy can be definitive. We can only say which effect we felt to be more obvious in our study. Finally, in our study as in virtually every other brain imaging study, DTI results cannot be confirmed by pathologic data because of the ethical issue of harm to the patient. As in other DTI tumor studies, pathologic correlation is very difficult to obtain because the neurosurgeon cannot resect surrounding soft tissue, such as WM tissue, for pathologic confirmation. In this study, we can only assume that extra-axial tumors, such as meningioma and metastases, will only cause edema or displacement of the adjacent WM tract. In contrast, intra-axial tumor, such as glioma and GBM, will always cause infiltration and destruction of the adjacent WM tracts. In summary, we defined an FA decrement (ΔFA%), which is the percentage change in FA that is associated with a lesion. Identifiable WM tracts adjacent to a tumor were selected and FA was calculated in a discrete anatomic region rather than in a mass of surrounding WM, so that the integrity of each affected WM tract could be assessed, preoperatively. Our results suggest that salvageable WM tracts can potentially be identified before surgery using this method. References [1] Field AS. Diffusion tensor imaging at the crossroads: fiber tracking meets tissue characterization in brain tumors. AJNR Am J Neuroradiol 2005;26:2168-9. [2] Helton KJ, Phillips NS, Khan RB, et al. Diffusion tensor imaging of tract involvement in children with pontine tumors. AJNR Am J Neuroradiol 2006;27:786-93. [3] Helton KJ, Weeks SJ, Phillips NS, et al. Diffusion tensor imaging of brainstem tumors: axonal degeneration of motor and sensory tracts. J Neurosurg Pediatrics 2008;1:270-6. [4] Inoue T, Ogasawara K, Beppu T, et al. Diffusion tensor imaging for preoperative evaluation of tumor grade in gliomas. Clin Neurol Neurosurg 2005;107:174-80. [5] Jellison BJ, Field AS, Medow J, et al. Diffusion tensor imaging of cerebral white matter: a pictorial review of physics, fiber tract anatomy, and tumor imaging patterns. AJNR Am J Neuroradiol 2004; 25:356-69. [6] Jena R, Prioce SJ, Baker C, et al. Diffusion tensor imaging: possible implications for radiotherapy treatment planning of patients with highgrade glioma. Clin Oncol (R Coll Radiol) 2005;17:581-90. [7] Khong PL, Kwong DLW, Chan GCF, et al. Diffusion-tensor imaging for the detection and quantification of treatment-induced white matter injury in children with medulloblastoma: a pilot study. AJNR 2003;24: 734-40. [8] Kinoshita M, Hashimoto N, Gogo T, et al. Fractional anisotropy and tumor cell density of the tumor core show positive correlation in diffusion tensor magnetic resonance imaging of malignant brain tumors. NeuroImage 2008 [on line].

P.S. Yen et al. / Surgical Neurology 72 (2009) 464–469 [9] Kinoshita M, Yamada K, Hashimoto N, et al. Fiber-tracking does not accurately estimate size of fiber bundle in pathological condition: initial neurosurgical experience using neuronavigation and subcortical white matter stimulation. Neuroimage 2005;25:424-9. [10] Lazar M, Alexander AL, Thottakara PJ, et al. White matter reorganization after surgical resection of brain tumors and vascular malformations. AJNR Am J Neuroradiol 2006;27:1258-71. [11] Melhem ER, Mori S, Mukundan G, et al. Diffusion tensor MR imaging of the brain and white matter tractography. AJR Am J Roentgenol 2002;178:3-16. [12] Nimsky C, Ganslandt O, Hastreiter P, et al. Intraoperative diffusiontensor MR imaging: shifting of white matter tracts during neurosurgical procedures—initial experience. Radiology 2005;234:218-25. [13] Pajevic S, Pierpaoli C. Color schemes to represent the orientation of anisotropic tissues from diffusion tensor data: application to white matter fiber tract mapping in the human brain. Magn Reson Med 1999; 42:526-40. [14] Papanikolaou N, Karampekios S, Papadaki E, et al. Fractional anisotropy and mean diffusivity measurements on normal human brain: comparison between low-and high-resolution diffusion tensor imaging sequences. Eur Radiol 2006;16:187-92. [15] Price SJ, Jena R, Burnet NG, et al. Improved delineation of glioma margins and regions of infiltration with the use of diffusion tensor imaging: an image-guided biopsy study. AJNR Am J Neuroradiol 2006;27:1969-74. [16] Provenzale JM, McGraw P, Mhatre P, et al. Peritumoral brain regions in gliomas and meningiomas: investigation with isotropic diffusionweighted MR imaging and diffusion-tensor MR imaging. Radiology 2004;232:451-60. [17] Skirboll SS, Ojemann GA, Berger MS, et al. Functional cortex and subcortical white matter located within gliomas. Neurosurgery 1996; 38:678-84. [18] Stieltjes B, Schluter M, Didinger B, et al. Diffusion tensor imaging in primary brain tumors: reproducible quantitative analysis of corpus callosum infiltration and contralateral involvement using a probabilistic mixture model. NeuroImage 2006;31:531-42. [19] Talos IF, Zou KH, Kikinis R, et al. Volumetric assessment of tumor infiltration of adjacent white matter based on anatomic MRI and diffusion tensor tractography. Acad Radiol 2007;14:431-6. [20] Toh CH, Castillo M, Wong AM, et al. Primary cerebral lymphoma and glioblastoma multiforme: differences in diffusion characteristics

[21]

[22]

[23]

[24]

[25]

[26]

469

evaluated with diffusion tensor imaging. AJNR Am J Neuroradiol 2008;29:471-5. Tropine A, Vucurevic G, Delani P, et al. Contribution of diffusion tensor imaging to delineation of gliomas and glioblastomas. J Magn Reson Imaging 2004;20:905-12. Tsuchiya K, Fujikawa A, Nakajima M, et al. Differentiation between solitary brain metastasis and high-grade glioma by diffusion tensor imaging. Brit J Radiol 2005;78:533-7. Verma R, Zacharaki EI, Ou Y, et al. Multiparametric tissue characterization of brain neoplasms and their recurrence using pattern classification of MR images. Acad Radiol 2008;15:966-77. Welzel T, Niethammer A, Mende U, et al. Diffusion tensor imaging screening of radiation-induced changes in the white matter after prophylactic cranial irradiation of patients with small cell lung cancer: first results of a prospective study. AJNR Am J Neuroradiol 2008;29: 379-83. Witwer BP, Moftakhar R, Hasan KM, et al. Diffusion-tensor imaging of white matter tracts in patients with cerebral neoplasm. J Neurosurg 2002;97:568-75. Zhang W, Olivi A, Hertig SJ, et al. Automated fiber tracking of human brain white matter using diffusion tensor imaging. NeuroImage 2008; 42:771-7.

Commentary Further assessment is needed to verify what the pathologic proof of DTI findings is. White matter changes can be found in any disease with DTI as fraction anisotrophy or apparent diffusion coefficient abnormality, but most of the articles do not investigate what it means. We need animal studies or autopsy data to better verify these findings. Noriko Salamon, MD Department of Radiology UCLA, Los Angeles CA 90095, USA