Differentiation of recurrent brain tumor versus radiation injury using diffusion tensor imaging in patients with new contrast-enhancing lesions

Abstract

Background and Purpose

The purpose of this study was to assess the use of diffusion tensor imaging (DTI) in the evaluation of new contrast-enhancing lesions and perilesional edema in patients previously treated for brain neoplasm in the differentiation of recurrent neoplasm from treatment-related injury.

Methods

Twenty-eight patients with new contrast-enhancing lesions and perilesional edema at the site of previously treated brain neoplasms were retrospectively reviewed.

Nine directional echoplanar DTIs with b=1000 s/mm² were obtained using a single-shot spin-echo echoplanar imaging. Standardized regions of interest were manually drawn in several regions. Mean apparent diffusion coefficient (ADC), fractional anisotropy (FA) and eigenvalue indices (λ_∥ and λ_⊥) and their ratios relative to the contralateral side were compared in patients with recurrent neoplasm versus patients with radiation injury, as established by histological examination or by clinical course, including long-term imaging studies and magnetic resonance spectroscopy.

Results

The ADC values in the contrast-enhancing lesions were significantly higher (P=.01) for the recurrence group (range=1.01×10⁻³ to 1.66×10⁻³ mm²/s; mean±S.D.=1.27±0.15) than for the nonrecurrence group (range=0.9×10⁻³ to 1.31×10⁻³ mm²/s; mean±S.D.=1.12±0.14).

The ADC ratios in the white matter tracts in perilesional edema trended higher (P=.09) in treatment-related injury than in recurrent neoplasm (mean±S.D.=1.85±0.30 vs. 1.60±0.27, respectively).

FA ratios were significantly higher in normal-appearing white matter (NAWM) tracts adjacent to the edema in the nonrecurrence group (mean±S.D.=0.89±0.15) than in those in the recurrence group (mean±S.D.=0.74±0.14; P=.03).

Both eigenvalue indices λ_∥ and λ_⊥ were significantly higher in contrast-enhancing lesions in the recurrence group than in those in the nonrecurrence group (P=.02). As well, both eigenvalue indices λ_∥ and λ_⊥ were significantly higher in perilesional edema than in normal white matter (P<.01 and P<.001, respectively) in both groups.

Conclusion

The assessment of diffusion properties, especially ADC values and ADC ratios, in contrast-enhancing lesions, perilesional edema and NAWM adjacent to the edema in the follow-up of new contrast-enhancing lesions at the site of previously treated brain neoplasms may add to the information obtained by other imaging techniques in the differentiation of radiation injury from tumor recurrence.

Introduction

Enhancing lesions that arise on routine follow-up brain magnetic resonance imaging (MRI) at the site of a previously identified and treated primary intracranial neoplasm may present a significant diagnostic dilemma. These lesions are typically subjected to radiation and/or chemotherapy and, in most instances, surgical resection. Many do not have specific imaging characteristics that enable the neuroradiologist to discriminate tumor recurrence from inflammatory or necrotic changes that result from treatment with radiation and/or chemotherapy [1]. Both recurrent tumors and treatment-related changes typically demonstrate enhancement with gadolinium. Therefore, often, the clinical course, brain biopsy or imaging over a lengthy follow-up interval is conclusive, not the specific imaging itself [1]. A noninvasive method for an earlier differentiation of recurrent tumor from treatment-related changes, when an abnormal contrast-enhancing lesion is first identified, would be invaluable. Certainly, other imaging MRI techniques, such as proton spectroscopy [2], [3], [4], [5] and MR perfusion [6], offer substantial potential in this regard. Lately, the use of positron emission tomography (PET) imaging especially has shown promising results in differentiating radiation injury from active neoplasm [7], [8]. Diffusion tensor imaging (DTI) is a more complex and complete form of diffusion imaging and has, for example, been able to demonstrate information on white matter tract altered by tumors [9], which can be valuable in presurgical planning, differentiating high-grade gliomas and evaluating the extent of cellular infiltration [10], [11]. In DTI, the directionality of water is probed by the application of diffusion sensitization gradients in multiple directions [12]. An appropriate mathematical combination of directional diffusion-weighted images provides quantitative measures of water diffusion for each voxel via the apparent diffusion coefficient (ADC), as well as the degree of diffusion directionality or anisotropy. In this work, the fractional anisotropy (FA) index was utilized. The tensor can be diagonalized such that only three nonzero elements (λ₁, λ₂ and λ₃) remain along the diagonal. These elements are known as eigenvalues. Each eigenvalue is associated with an eigenvector (ε₁, ε₂ and ε₃), where the largest of the three eigenvalues (λ₁) corresponds to the eigenvector ε₁ and describes the principal direction of diffusion at that point. Studies using more sophisticated methods for the evaluation of tumors, such as analyses of the role of different eigenvectors, have demonstrated that the value of the major eigenvector of diffusion reflecting diffusivity in the longitudinal direction was significantly lower in the white matter surrounding the glioma than that in the white matter surrounding metastasis, even when the anisotropy showed no difference [13]. Studies like this and other suggest that more sophisticated approaches might yield more information than just mean diffusivity and anisotropy measurements. The incorporation of DTI parameters into a decision rule for differentiating recurrent neoplasms from posttreatment changes represents such an approach. However, prior to the development of a decision rule, we need to understand the distribution of the values of these parameters. Therefore, the purpose of this work is to examine the ability of DTI to differentiate recurrent neoplasms from treatment-related injuries. To that end, we compared standard DTI parameters (ADC and FA values, ADC and FA ratios) and the values and ratios of the eigenvalue indices λ_∥ (principal eigenvalue) and λ_⊥ (mean of eigenvalues perpendicular to λ_∥) in three clinically relevant regions commonly visualized in imaging examinations, namely, in contrast-enhancing lesions, tracts in perilesional edema and tracts in the surrounding normal-appearing white matter (NAWM).