Utilizing Diffusion Tensor Imaging to Differentiate High-Grade Gliomas and Solitary Brain Metastases

Abstract

Background Brain tumors, encompassing a spectrum of neoplastic disorders, significantly impact patient morbidity and mortality. Distinguishing between high-grade gliomas (HGGs) and solitary brain metastases (SBMs) is crucial for tailored clinical management. Conventional structural magnetic resonance imaging (MRI) faces challenges in this differentiation, leading to the exploration of advanced imaging modalities such as diffusion tensor imaging (DTI).

Materials and Methods In this prospective study, 41 patients with solitary enhancing brain lesions underwent total or subtotal resection, confirmed by histopathology. Imaging involved a 3-Tesla MRI scanner, and DTI data were analyzed for metrics including mean diffusivity, fractional anisotropy (FA), axial diffusivity, radial diffusivity, as well as planar, spherical, and linear (CL) anisotropy coefficients.

Results Peritumoral FA and CL exhibited significant differences (p = 0.0217 and p = 0.039, respectively) between HGG and SBM. The area under the curve for peritumoral FA and CL in differentiating HGG and SBM were 0.2791 and 0.6984, respectively. No significant differences were observed in the other diffusion metrics.

Conclusion This study contributes to understanding DTI-derived metrics for HGG and SBM differentiation. Peritumoral FA and CL show promise as potential discriminators, offering insights for enhanced clinical decision-making and treatment planning in brain tumor patients. Future research with larger cohorts and advanced diffusion imaging techniques could further refine these findings.

Keywords

diffusion tensor imaging - high-grade glioma - solitary brain metastases - mean diffusivity and fractional anisotropy

Introduction

Brain tumors represent a diverse and complex group of neoplastic disorders that can significantly impact both patient morbidity and mortality.[1] Among the various types of brain tumors, differentiating between them is of paramount importance due to their distinct clinical management and prognostic implications.[2] High-grade gliomas (HGGs), characterized by their aggressive growth and infiltrative nature within the brain parenchyma, present a unique diagnostic challenge.[3] Conversely, solitary brain metastases (SBMs), originating from primary tumors outside the central nervous system, demand precise identification to guide treatment decisions and determine prognosis.[4]

Conventional structural magnetic resonance imaging (MRI) techniques have provided crucial insights into the anatomical characteristics of brain tumors. However, accurately distinguishing between HGG and SBM remains challenging due to overlapping radiological features such as contrast enhancement and peritumoral edema. The limited specificity of conventional MRI has spurred the exploration of advanced imaging modalities that can provide deeper insights into the underlying tissue microstructure and pathophysiology of these tumors, including diffusion-weighted imaging, diffusion tensor imaging (DTI), dynamic susceptibility imaging, dynamic contrast enhancement, and MR spectroscopy.[5]

DTI has emerged as a powerful neuroimaging technique that offers unique information about tissue microarchitecture by probing the diffusion of water molecules within biological tissues. The diffusion properties of water are influenced by barriers such as cell membranes, extracellular matrix, and fiber tracts. As a result, DTI can provide valuable information about tissue integrity, cellular density, and fiber orientation. In brain tumor research, DTI has shown promise in elucidating the underlying differences in tissue microstructure between different tumor types.[6]

The ability of DTI to quantify metrics such as mean diffusivity (MD) and fractional anisotropy (FA) has raised significant interest in its potential to discriminate between HGG and SBM. MD reflects the overall magnitude of water diffusion and can be influenced by factors such as cellularity and tissue disruption. FA, on the other hand, provides information about the directionality and coherence of water diffusion, thereby offering insights into tissue microarchitecture and organization. These metrics have demonstrated sensitivity to changes in tissue composition and disruption caused by pathological processes.[7] [8] Using these diffusion metrics, varying success has been achieved in distinguishing between “infiltrative edema” of HGG and “vasogenic edema” of metastases. There also seem to be no significant differences in the intratumoral metrics of these two entities.[9] [10] [11] [12]

Peritumoral vasogenic edema exhibits notably elevated MD compared with infiltrative edema, primarily attributed to the significant increase in water accumulation in contrast to the cellular infiltration seen in glioblastoma. On the other hand, differences in peritumoral FA between the two pathologies are not consistently observed. This discrepancy might arise from the varying proportions of normal white and gray matter, intracellular water, and tumor cell infiltration present within infiltrative edema.[13]

In addition to MD and FA, DTI provides several other scalar metrics like axial diffusivity (AD), radial diffusivity (RD), as well as planar (CP), spherical (CS), and linear (CL) anisotropy coefficient.[8] These metrics have been studied in glioma grading,[14] [15] but have not been utilized to differentiate HGG and SBM.

This study aims to contribute to the growing body of research investigating the utility of DTI-derived metrics in differentiating between HGG and SBM. By evaluating these values within both intratumoral and peritumoral regions, we aim to assess whether there are any significant differences between the two tumor types. The potential implications of our findings extend to enhancing clinical decision-making, prognostic assessment, and treatment planning for patients with brain tumors. Furthermore, a deeper understanding of the distinct DTI profiles associated with HGG and SBM may shed light on the underlying biological mechanisms driving their behavior and progression.

Materials and Methods

The study was commenced after getting clearance from the Institutional Ethical Committee. We prospectively enrolled 41 patients over 2 years, presenting with a solitary enhancing lesion on postcontrast T1 images. Patients with a previous history of cranial surgery, chemotherapy, or radiotherapy, general contraindications for MRI, contrast allergy, and multiple parenchymal enhancing lesions were excluded from the study.

All the patients underwent total or subtotal resection, and the diagnosis was confirmed on histopathology.

Imaging technique and data analysis: MRI was performed using a 3-Tesla scanner (GE Architect) using a Head 48 channel coil, the T1 (TR/TE, 7700/1000 ms), T2 (TR/TE, 5742/1000 ms), and fluid-attenuated inversion recovery (TR/TE/TI, 7500/120/1000 ms) sequences with matrix 256 × 256, field of view (FOV) 24 × 24 (frequency and phase FOV) and slice thickness of 1.8 mm were obtained. Post-intravenous contrast T1 image study was done using gadolinium-based contrast (gadobuterol) (0.5 mL/kg [0.1 mmol/kg] body weight with the maximum dose of 10 mL at a flow rate of ∼2 mL/s).

DTI data were obtained using a single-shot echo planar imaging sequence (TR/TE 3941/1000 ms) with parallel imaging that includes two strong encoding gradients as described by Stejskal-Tanner.[16] Diffusion gradients were applied along 50 directions, using a b-value of 0 and 1,000 second mm². The frequency FOV was 26, the phase FOV was 1 and a data matrix of 128 × 128 was used. Thirty-eight slices were obtained, a thickness of 3 mm, with no gap, and a total scan duration of approximately 3 to 4 minutes.

These gradients are balanced concerning stationary protons, induce a diffusion-dependent phase dispersion, and result in signal loss in areas with higher diffusion. The diffusion sensitivity of the pulse sequence is expressed as the b-value in units s/m² and is related to the duration and strength of the diffusion encoding gradient. Measurements including two b-values (for example, b = 0 second mm² and b= 1,000 second mm²) allow for plotting the logarithmic signal intensity versus the b-value for a region-of-interest (ROI) or a volume element (voxel).

The data was transferred to the GE workstation for postprocessing, and the images were analyzed using software with automated generation of diffusion maps. After using a correction algorithm to compensate for head motion and image distortion due to the eddy currents, the FA, MD, AD, RD, CL, CS, and CP diffusion tensor maps were generated.

In the context of the DTI model, FA indicates the degree of asymmetry in diffusion within a voxel, with values theoretically ranging between 0 (completely isotropic diffusion) and 1 (completely anisotropic diffusion). MD indicates how much water molecules are spreading out within a voxel, giving an overall measure of molecular dispersion. AD and RD refer to the average rates of diffusion of water molecules in specific directions. AD measures the average diffusion parallel to the primary direction indicated by the largest eigenvalue of the diffusion tensor, while RD quantifies the average diffusion perpendicular to this primary direction.[17]

CL, CP, and CS provide additional information about the characteristics of the diffusion ellipsoid's shape. CL indicates how elongated or linear the ellipsoid is, with a focus on diffusion occurring predominantly along the axis corresponding to the largest eigenvalue. CP, on the other hand, quantifies how the ellipsoid is flattened or planar, emphasizing diffusion primarily within the plane formed by the two eigenvectors associated with the two largest eigenvalues. Lastly, CS measures the degree of spherical shape or isotropic diffusion within the ellipsoid.[17]

Before the evaluation of the diffusion maps, conventional MRI sequences were assessed for lesion localization, characterization, and enhancement pattern. For quantitative evaluation, we manually placed four circular ROIs (ROIs: 5 mm²) within the enhancing component of the lesion and in the region of peritumoral edema (within 1 cm from the enhancing lesion) in an orthogonal orientation on postcontrast T1 images and transferred to the DTI postprocessed maps. The ROIs were selected manually by an experienced radiologist with more than 4 years of experience in DTI analysis (S.R.K.), aiming to maintain consistency. Each ROI was placed in the most enhanced region to ensure consistency and focus on the areas of maximal tumor cellularity and enhancement.

Means of all the diffusion metrics across four ROIs of intratumoral and peritumoral regions were obtained from DTI postprocessed images and not normalized to cerebral white matter.

Statistical Analysis

The Statistical Package for the Social Sciences, version 23, was used to conduct statistical analysis (SPSS-23, IBM, Chicago, Illinois, United States).

The Mann–Whitney test was used to compare the diffusion metrics values between the two groups. Statistical significance was defined as a p-value of < 0.05.

The results of the Mann–Whitney U test are tabulated in [Table 1].

Table 1 Results of Mann–Whitney U test

DTI metric	Group	50th percentile	p-Value (significant p-Value < 0.05)
Intratumoral FA	SBM HGG	0.171 0.166	0.277
Intratumoral MD	SBM HGG	1.301 1.092	0.5824
Intratumoral AD	SBM HGG	0.920 0.981	0.4831
Intratumoral RD	SBM HGG	0.983 1.012	0.9233
Intratumoral CP	SBM HGG	0.011 0.037	0.7415
Intratumoral CS	SBM HGG	0.996 0.992	0.8152
Intratumoral CL	SBM HGG	0.002 0.001	0.9015
Peritumoral FA	SBM HGG	0.112 0.216	0.0217 (significant)
Peritumoral MD	SBM HGG	1.508 1.246	0.2264
Peritumoral AD	SBM HGG	1.68 1.47	0.0857
Peritumoral RD	SBM HGG	1.68 1.46	0.2714
Peritumoral CP	SBM HGG	0.02 0.009	0.1824
Peritumoral CS	SBM HGG	1.018 0.997	0.8690
Peritumoral CL	SBM HGG	0.028 0.018	0.0392 (significant)

Abbreviations: AD, axial diffusivity; CL, linear coefficient; CP, planar coefficient; CS, spherical coefficient; DTI, diffusion tensor imaging; FA, fractional anisotropy; HGG, high-grade glioma; MD, mean diffusivity; RD, radial diffusivity; SBM, solitary brain metastasis.

Results

The study population consisted of 27 HGGs (grade III and IV) (19 men, 8 women, age range 40–75 years, mean age 55 ± 11.2 [standard deviation]) and 14 SBMs (6 men, 8 women, age range 35–80 years, mean age 58 ± 14.1 [standard deviation]). Of the 14 patients with SBM, 9 patients had primary lung malignancy, 3 had colon adenocarcinoma, and 2 had breast carcinoma.

There is a significant difference in peritumoral FA (p = 0.0217) and peritumoral CL (p = 0.039) between the two groups when analyzed using the Mann–Whitney U test. The median value was higher in the HGG group compared with SBM in the case of peritumoral FA, while it was higher in SBM compared with HGG in the case of peritumoral CL. At the receiver operating characteristic curve, the area under the curve (AUC) of peritumoral FA used to differentiate HGG and SBM was 0.2791. Similarly, the AUC of peritumoral CL was 0.6984. Box plots describing peritumoral FA and CL between SBMs (group 1) and HGGs (group 2) are provided in [Fig. 1].

Fig. 1 (A, B) In the graphical representation, the interquartile range between the 25th and 75th percentiles is depicted by boxes. The medians are denoted by the central lines within these boxes. The vertical bars extending from the boxes, referred to as whiskers, illustrate the data range excluding outliers. Group 1: solitary brain metastasis and group 2: high-grade glioma. Print resolution: 600 dpi. Size (Px): W: 1138 H: 414. Print size: 4.82 * 1.75 cm.

There were no significant statistical differences in the other five diffusion metrics between the two groups. Representative cases of HGG and SBM are depicted in [Figs. 2] and [3].

Fig. 2 (A–D) A 45-year-old female presented with headache and seizures. (A) Postcontrast axial T1 image shows an enhancing lesion with peritumoral edema. (B) Axial fluid-attenuated inversion recovery (FLAIR) image shows the peritumoral edema associated with the lesion. Representative regions of interest (ROIs) are placed in the enhancing area and peritumoral area in orthogonal orientation. (C and D) Axial fractional anisotropy and mean diffusivity map of the lesion with ROIs. Histopathological diagnosis: Anaplastic astrocytoma World Health Organization (WHO) grade III. Print resolution: 600 dpi. Size (Px): W: 3007 H: 921. Print size: 5.01 * 1.54 inches.

Fig. 3 (A–C) A 55-year-old male presented with right-sided weakness. (A) Postcontrast axial T1 image shows an enhancing lesion with central nonenhancing areas and peritumoral edema. Representative regions of interest (ROIs) are placed in the enhancing area and peritumoral area in orthogonal orientation. (B) Axial fluid-attenuated inversion recovery (FLAIR) image shows the peritumoral edema associated with the lesion. (C and D) Axial fractional anisotropy and mean diffusivity map of the lesion with ROIs. Histopathological diagnosis: Metastasis from primary lung malignancy. Print resolution: 600 dpi. Size (Px): W: 3190 H: 967. Print size: 5.32 * 1.61 inches.

Discussion

The ability of the DTI-generated parameters from the lesion and perilesional edema to differentiate between several groups of brain tumors has been statistically evaluated in this study. This was predicated on the fact that perilesional edema in brain lesions might vary in composition. The majority of benign lesions exhibit vasogenic perilesional edema; however, gliomas also exhibit infiltrating cells. In this investigation, variations in the DTI parameters were assessed.

HGGs have higher perilesional FA because glioblastomas have more peritumoral infiltration than metastases and produce a more tumor-specific extracellular matrix, which results in increased anisotropy, as found in our study.[14] Extracellular water is a key indicator of MD because metastatic lesions express high levels of vascular endothelial growth factor, which promotes vascular permeability. A larger perilesional MD in metastases than in HGGs can be attributed to the latter cause; however, it was found to be statistically insignificant.[10]

In our study, peritumoral CL tensor was higher in SBM when compared with HGG. This finding was contrary to previous studies, where peritumoral CL was higher in HGG.[18] [19] Varying compositions of edema, tumor infiltration, normal gray/white matter, and the adjacent tracts in the immediate peritumoral region complicate the identification of the precise factors responsible for the variations in these metrics.[5]

Wang et al[20] studied pure isotropic diffusion (p), pure anisotropic diffusion (q), and the total magnitude of the diffusion tensor (L) and FA parameters in glioblastoma multiforme (GBM) and brain metastases, revealing significantly reduced q and FA of the gross tumor portion, and q in the peritumoral margin in the GBM group. They also concluded that none of these metrics can distinguish between these two groups by itself.

Several studies have concluded that the enhancing region of GBM had higher FA, CP, and CL, and lower CS as compared with metastases.[19] [21] Regression coefficients of RD and AD were significantly higher in the tumor-infiltrated edema, as compared with pure vasogenic edema.[22] As MD and RD are inversely related to the degree of tumor cellularity, lower intratumoral MD and RD have been found in HGG when compared with SBM.[23]

A meta-analysis conducted by Jiang et al revealed higher FA and lower MD in the peritumoral areas of HGGs when compared with metastasis.[10] Another meta-analysis by Suh et al revealed a pooled sensitivity of 77.0% (95% confidence interval [CI], 62.3–87.1%) and a pooled specificity of 80.3% (95% CI, 73.5–85.7%) for differentiating these two entities based on DTI, with peritumoral MD showing the highest sensitivity and specificity.[9]

Newer diffusion imaging techniques such as neurite orientation dispersion and density imaging (NODDI), mean apparent propagator MRI, and diffusion kurtosis imaging (DKI) have been studied, with NODDI demonstrating the highest performance.[23]

There are several drawbacks in this study. The first is the relatively small sample size in this study, thereby limiting the extrapolation of our conclusions. The second limitation is the subjective assessment and manual placement of ROI. Since we used a predefined size for ROI, placed them within 1 cm of the enhancing region, utilized multiple ROIs, and used the mean value, this limitation was negated to a certain extent. However, the possibility of exclusion of infiltrative cells in the peritumoral edema from measurement cannot be ruled out.[24] This limitation can be negated to a certain extent by using semiautomated methods to obtain a greater sampling area,[25] and by using an interrater reliability analysis.[12] Though histopathological diagnosis was available for all the cases, the peritumoral area was not assessed during biopsy/surgery, though several studies have correlated DTI and pathological findings.[26] [27] Utilizing average parameter values in frequently heterogeneous areas may not consistently provide the most accurate representation of anisotropy within the region.[19] Further analysis of SBM based on their primary origin could not be assessed due to limited data.

Conclusion

Our study highlights the potential utility of DTI-derived metrics in distinguishing between HGGs and SBMs. Specifically, peritumoral FA and linear anisotropy coefficient (CL) demonstrated significant differences between the two tumor types. The observed higher peritumoral FA in HGG and elevated peritumoral CL in SBM suggest distinct microstructural characteristics in these regions. While these findings contribute to the growing body of research, the study acknowledges limitations such as a relatively small sample size and subjective ROI placement. Nonetheless, the insights gained from this investigation could have implications for refining clinical decision-making, prognostic assessment, and treatment planning for patients with brain tumors. Future research endeavors, encompassing larger cohorts and advanced diffusion imaging techniques, may further elucidate and validate these findings, paving the way for more precise characterization and management of these diverse brain neoplastic disorders.

Conflict of Interest

None declared.

Acknowledgments

We thank the Departments of Radiology, Neurology, and Neurosurgery for their support during the study.

Ethical Approval

IEC Study Ref No- 350/2020: The study titled “DTI in Distinguishing Gliomas vs. Brain Metastases” has been approved on 27/11/2020 till the end of study.

• References

• 1 Ostrom QT, Francis SS, Barnholtz-Sloan JS. Epidemiology of brain and other CNS tumors. Curr Neurol Neurosci Rep 2021; 21 (12) 68
• 2 Rivera M, Norman S, Sehgal R, Juthani R. Updates on surgical management and advances for brain tumors. Curr Oncol Rep 2021; 23 (03) 35
• 3 Nayak L, Reardon DA. High-grade gliomas. Continuum (Minneap Minn) 2017; 23 (6, Neuro-oncology): 1548-1563
• 4 Boire A, Brastianos PK, Garzia L, Valiente M. Brain metastasis. Nat Rev Cancer 2020; 20 (01) 4-11
• 5 Fordham AJ, Hacherl CC, Patel N. et al. Differentiating glioblastomas from solitary brain metastases: an update on the current literature of advanced imaging modalities. Cancers (Basel) 2021; 13 (12) 2960
• 6 Jellison BJ, Field AS, Medow J, Lazar M, Salamat MS, Alexander 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 (03) 356-369
• 7 Pierpaoli C, Jezzard P, Basser PJ, Barnett A, Di Chiro G. Diffusion tensor MR imaging of the human brain. Radiology 1996; 201: 637–648. Cerebrovasc Dis 2005; 20: 310-318
• 8 O'Donnell LJ, Westin CF. An introduction to diffusion tensor image analysis. Neurosurg Clin N Am 2011; 22 (02) 185-196 , viii
• 9 Suh CH, Kim HS, Jung SC, Kim SJ. Diffusion-weighted imaging and diffusion tensor imaging for differentiating high-grade glioma from solitary brain metastasis: a systematic review and meta-analysis. AJNR Am J Neuroradiol 2018; 39 (07) 1208-1214
• 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 (11) e112550
• 11 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 (03) 477-483
• 12 Wang S, Kim SJ, Poptani H. et al. Diagnostic utility of diffusion tensor imaging in differentiating glioblastomas from brain metastases. AJNR Am J Neuroradiol 2014; 35 (05) 928-934
• 13 Lee EJ, Ahn KJ, Lee EK, Lee YS, Kim DB. Potential role of advanced MRI techniques for the peritumoural region in differentiating glioblastoma multiforme and solitary metastatic lesions. Clin Radiol 2013; 68 (12) e689-e697
• 14 El-Serougy L, Abdel Razek AAK, Ezzat A, Eldawoody H, El-Morsy A. Assessment of diffusion tensor imaging metrics in differentiating low-grade from high-grade gliomas. Neuroradiol J 2016; 29 (05) 400-407
• 15 Jiang L, Xiao CY, Xu Q. et al. Analysis of DTI-derived tensor metrics in differential diagnosis between low-grade and high-grade gliomas. Front Aging Neurosci 2017; 9: 271
• 16 Borkowski K, Krzyżak AT. The generalized Stejskal-Tanner equation for non-uniform magnetic field gradients. J Magn Reson 2018; 296: 23-28
• 17 Westin CF, Maier SE, Mamata H, Nabavi A, Jolesz FA, Kikinis R. Processing and visualization for diffusion tensor MRI. Med Image Anal 2002; 6 (02) 93-108
• 18 Wang S, Kim S, Chawla S. et al. Differentiation between glioblastomas and solitary brain metastases using diffusion tensor imaging. Neuroimage 2009; 44 (03) 653-660
• 19 Wang S, Kim S, Chawla S. et al. Differentiation between glioblastomas, solitary brain metastases, and primary cerebral lymphomas using diffusion tensor and dynamic susceptibility contrast-enhanced MR imaging. AJNR Am J Neuroradiol 2011; 32 (03) 507-514
• 20 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 (01) 203-208
• 21 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 (09) 1646-1651
• 22 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 (08) 1366-1370
• 23 Mao J, Zeng W, Zhang Q. et al. Differentiation between high-grade gliomas and solitary brain metastases: a comparison of five diffusion-weighted MRI models. BMC Med Imaging 2020; 20 (01) 124
• 24 van Westen D, Lätt J, Englund E, Brockstedt S, Larsson EM. Tumor extension in high-grade gliomas assessed with diffusion magnetic resonance imaging: values and lesion-to-brain ratios of apparent diffusion coefficient and fractional anisotropy. Acta Radiol 2006; 47 (03) 311-319
• 25 Holly KS, Barker BJ, Murcia D. et al. High-grade gliomas exhibit higher peritumoral fractional anisotropy and lower mean diffusivity than intracranial metastases. Front Surg 2017; 4: 18
• 26 Price SJ, Burnet NG, Donovan T. et al. Diffusion tensor imaging of brain tumours at 3T: a potential tool for assessing white matter tract invasion?. Clin Radiol 2003; 58 (06) 455-462
• 27 Lu S, Ahn D, Johnson G, Law M, Zagzag D, Grossman RI. Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology 2004; 232 (01) 221-228

Publication History

Article published online:
13 January 2025