European Journal of Cardiovascular Medicine

Brain tumours comprise a heterogeneous group of primary and secondary neoplasms that differ widely in cellular origin, biological behaviour, and prognosis. Contemporary classification (WHO CNS5) integrates histology with molecular markers (e.g., IDH mutation, 1p/19q codeletion, EGFR alterations) and this integrated taxonomy has reshaped diagnostic algorithms, prognostication, and therapeutic decision-making. Advanced MRI methods — including diffusion, perfusion, and magnetic resonance spectroscopy (MRS) — provide complementary, non-invasive information about tumour cellularity, vascularity, and metabolism that helps refine diagnosis, grade prediction, biopsy targeting, and treatment monitoring. [1,2].

Tumour-specific MRI and MRS Features of Brain Tumours

Diffuse Gliomas (Astrocytomas, Oligodendrogliomas, Glioblastomas)

Diffuse gliomas are the most common primary malignant brain tumours in adults and are classified according to molecular markers such as IDH mutation and 1p/19q codeletion, which influence prognosis and treatment. On MRI, low-grade gliomas typically appear as non-enhancing T2/FLAIR hyperintense lesions with minimal mass effect, whereas high-grade gliomas demonstrate irregular contrast enhancement, necrosis, and extensive peritumoural oedema. [1] MRS shows elevated choline (Cho) due to increased membrane turnover and reduced N-acetylaspartate (NAA) indicating neuronal loss. High-grade tumours exhibit markedly increased Cho/NAA and Cho/Cr ratios with lipid–lactate peaks, while low-grade gliomas show milder metabolic abnormalities. Detection of 2-hydroxyglutarate (2HG) can serve as a non-invasive marker of IDH-mutant gliomas. [3,4]

Meningiomas
Meningiomas are common extra-axial tumours, usually benign, characterized on MRI by strong homogeneous enhancement, a dural tail, and occasional hyperostosis. On MRS, they typically demonstrate absent or markedly reduced NAA, elevated Cho, and a characteristic alanine peak, reflecting their non-neuronal origin. Although these metabolic findings aid diagnosis, overlap with other extra-axial lesions and technical limitations reduce specificity; therefore, MRS serves as a useful adjunct to conventional MRI. [2,3]

Metastases and Primary CNS Lymphoma

Brain metastases commonly arise from lung, breast, melanoma, renal, or colorectal malignancies and usually present as multiple enhancing lesions at the grey–white matter junction with surrounding vasogenic oedema.[2] MRS reveals increased Cho and reduced NAA within the tumour; however, the surrounding oedema generally shows preserved neuronal metabolites, unlike infiltrative gliomas where metabolic abnormalities extend beyond tumour margins.[5,6] Primary CNS lymphoma typically appears as a homogeneously enhancing deep-seated lesion with marked diffusion restriction. Spectroscopy demonstrates elevated Cho, reduced NAA, and occasional lipid peaks, helping differentiate it from glioblastoma and metastases when combined with advanced MRI techniques. [3]

Sellar and Paediatric Tumours

MRS has a more limited role in sellar lesions because of technical challenges related to small lesion size and adjacent bone. Nevertheless, pituitary adenomas may demonstrate altered metabolite profiles, while craniopharyngiomas often show lipid or cholesterol peaks. In paediatric embryonal tumours such as medulloblastoma, MRS typically reveals high Cho, low NAA, and characteristic lipid or taurine peaks, aiding tumour characterization and treatment monitoring. [7,8]

Role of MRI in Brain Tumour Evaluation

MRI remains the cornerstone of brain tumour imaging because of its excellent soft-tissue contrast and multiparametric capabilities. Conventional MRI combined with advanced techniques such as diffusion imaging, perfusion imaging, and MRS provides comprehensive structural, physiological, and metabolic information. This multimodal approach improves tumour characterization, grading, surgical planning, treatment response assessment, and non-invasive prediction of tumour biology and molecular status. [9]

Introduction to Magnetic Resonance Spectroscopy (MRS)

Magnetic Resonance Spectroscopy (MRS) is an MRI-based, non-invasive method that acquires a frequency-domain spectrum from a localized brain region and reports resonances from endogenous metabolites (most commonly ^1H). Unlike conventional MRI (anatomic contrast), MRS reports tissue biochemistry — concentrations or relative levels of metabolites that reflect membrane turnover, neuronal integrity, energy metabolism, anaerobic glycolysis, necrosis, and specific oncometabolites. [2]

Magnetic Resonance Spectroscopy (MRS): Metabolic profiling and Molecular surrogates

Proton MRS noninvasively measures metabolites such as choline (membrane turnover), N-acetyl aspartate (neuronal integrity), creatine (energy metabolism), lactate and lipids (anaerobic metabolism/necrosis), and 2-hydroxyglutarate for IDH mutation. Tumours typically show elevated choline and reduced NAA with increased Cho/NAA and Cho/Cr ratios. MRS aids tumour grading, biopsy targeting, molecular characterization, and differentiation from abscess. MRS helps improve confidence when conventional imaging is equivocal. [2]

Place of Study: The study was conducted at Tezpur Medical College & Hospital. Duration of Study: The study was conducted over a period of one year. Study Design: Institutional based Cross-sectional descriptive study. Study Population: Patients included in the study were those who were clinically referred for MRI Brain in the Department of Radiodiagnosis at Tezpur Medical College & Hospital. Sample Size: 68 Inclusion Criteria • Patients of both sexes and all age groups. • Patients who have given informed consent. • Patients with high clinical suspicion / Diagnosed case of Brain tumours. Exclusion Criteria • Patients with infective etiology, intracranial bleed, demyelinating lesions, ischaemic lesions and other Neurological Conditions such as neurodegenerative diseases, are excluded to avoid confounding spectroscopic findings. • Contraindication to MRI study such as patients with pacemakers, metallic implants and aneurysmal clips • Patients having Claustrophobia/who don’t wish to participate in the study/Inability to Provide Informed Consent. • Unconfirmed Diagnosis. • Patients who have undergone prior treatment may be excluded to minimize the influence of treatment-related changes on spectroscopic findings. • Severe Medical Comorbidities: Participants with severe medical conditions that could affect brain metabolism or spectroscopic measurements, such as end-stage renal disease or severe hepatic dysfunction, may be excluded. • Pregnancy. The institutional ethical committee approval was obtained properly before the commencement of study.

DEMOGRAPHIC CHARACTERISTICS OF STUDY POPULATION

Table 1: Age and Sex Distribution Among Study Subjects (n = 68)

The mean age of the study population was 40.8 ± 22.6 years (range: 3–88 years).

Majority of patients were in the 41–60 years age group (33.8%), with peak incidence in the 51–60 years group (17.6%). Paediatric patients (≤20 years) constituted 19.1% of cases. A male predominance was observed, accounting for 61.8% of the study population.

DISTRIBUTION OF BRAIN TUMOUR TYPES

Table 2: Distribution of Study Population based on Brain Tumour Diagnoses

High-grade gliomas were the most common tumour type (29.4%), followed by metastatic    tumours and meningiomas.

Table 3: Mean Age Distribution of Patients According to Tumour Type

The bar graph illustrates the highest mean age was observed in high-grade glioma, followed by acoustic schwannoma and metastatic brain tumours, indicating their predominance in older patients. While medulloblastoma exhibits the lowest mean age, consistent with its predominance in the paediatric population.

METABOLITE RATIO ANALYSIS

Table 4: Mean ± SD of Metabolite Ratios According to Diagnosis

Significant differences in metabolite ratios were observed across different tumour types.

• High-grade gliomas demonstrated markedly elevated Cho/NAA (3.15 ± 0.78) and Cho/Cr (2.92 ± 0.63) ratios with significantly reduced NAA/Cr (0.42 ± 0.18).
• Low-grade gliomas showed lower Cho/NAA (1.32 ± 0.28) and Cho/Cr (1.41 ± 0.31) with relatively preserved NAA/Cr (1.18 ± 0.27).
• Metastatic tumours showed moderately elevated Cho/NAA ratios (2.23 ± 0.68) with reduced NAA/Cr.

The differences in metabolite ratios between tumour groups were statistically significant (ANOVA, p < 0.001).

This figure illustrates the distribution of Cho/NAA ratios among various brain tumour types. One-way ANOVA demonstrated a statistically significant difference in Cho/NAA ratios across tumour groups (F = 32.754, p < 0.001), confirming strong discriminatory capability of this metabolic parameter.

This figure demonstrates the distribution of Cho/Cr ratios among different tumour categories.

This figure shows the distribution of NAA/Cr ratios across tumour types. High grade glioma, medulloblastoma, and metastatic tumours demonstrate reduced NAA/Cr ratios, reflecting neuronal loss and tumour infiltration.

The difference in NAA/Cr ratios across tumour groups was statistically significant (F = 18.129, p < 0.001), supporting its inverse correlation with tumour aggressiveness.

CORRELATION OF MRS WITH HISTOPATHOLOGY

• Positive (High-grade):
• Cho/NAA > 2.0 OR
• Cho/NAA 1.8–2.0 with lipid-lactate ≥2

Table 6: Correlation of MRS findings with Histopathology (n = 68)

The following diagnostic indices were calculated:

These findings demonstrate a good concordance between MRS interpretation and histopathological diagnosis, with some overlap between tumour categories.

Table 7: Correlation of Metabolite Ratios with Tumour Grade (Pearson correlation analysis)

The bar diagram illustrates metabolite ratios Cho/NAA (r = +0.72) and Cho/Cr (r = +0.68) demonstrate strong positive correlations, indicating increased membrane turnover and cellular proliferation with higher tumour grade.

In contrast, NAA/Cr (r = –0.61) shows a negative correlation, reflecting progressive neuronal loss with increasing tumour grade.

Table 8: ROC Curve Analysis of Metabolite Ratios for Differentiating High-Grade from Low-Grade Tumours

Receiver Operating Characteristic (ROC) curve analysis demonstrated that all metabolite ratios showed statistically significant diagnostic performance. Cho/NAA demonstrated the highest AUC (0.68, p < 0.01), followed by Cho/Cr (AUC = 0.66, p < 0.05) and NAA/Cr (AUC = 0.65, p < 0.05).

The Area Under the Curve (AUC) values for all metabolite ratios were statistically significant (p < 0.05), indicating moderate discriminatory ability of MR Spectroscopy in differentiating high-grade from low-grade brain tumours.

Magnetic Resonance Spectroscopy (MRS) provides non-invasive metabolic characterization of brain tumours and complements conventional MRI in tumour grading and differentiation. The present study evaluated the role of MRS in differentiating tumour types and grades using metabolite ratios and correlated these findings with histopathology. The results demonstrated statistically significant differences in metabolite profiles across tumour categories, with strong diagnostic performance of MRS.

Demographic Profile

In the present study, the mean age of patients was 40.8 ± 22.6 years, with the majority belonging to the 41–60 years age group (33.8%). A male predominance (61.8%) was observed. These findings are comparable with previously published studies, which report higher incidence of brain tumours in middle-aged adults with slight male preponderance. The inclusion of paediatric cases (19.1%) reflects the heterogeneous nature of brain tumours across age groups.

Gliomas (High-Grade vs Low-Grade)

High-grade gliomas constitute 29.4% of cases and demonstrate markedly elevated Cho/NAA (3.15 ± 0.78) and Cho/Cr (2.92 ± 0.63) ratios with significantly reduced NAA/Cr (0.42 ± 0.18), whereas low-grade gliomas showed lower Cho/NAA (1.32 ± 0.28) and relatively preserved NAA/Cr (1.18 ± 0.27), with statistically significant differences (p < 0.001).

These findings are consistent with Li X et al. (2020) who demonstrated progressive elevation of choline ratios with increasing tumour grade.[4] Galijasevic M et al. (2022) also demonstrated that Cho/NAA and Cho/Cr ratios reliably differentiate high-grade from low-grade gliomas. The elevated choline reflects increased cellular proliferation and membrane turnover [8]. Furthermore, Shakir TM et al. (2022) suggested a Cho/NAA threshold of >2.0 for high-grade gliomas, which was exceeded in the majority of high-grade cases in our cohort[10]. These findings are consistent with the foundational work by Law et al. (2003) and Kim et al. (2006), who established metabolite ratios as reliable biomarkers for glioma grading[11,12].

Brain Metastases

Metastatic tumours demonstrated moderately elevated Cho/NAA (2.23 ± 0.68) with reduced NAA/Cr and frequent lipid–lactate peaks. These values were lower than those observed in high-grade gliomas. This aligns with Claus EB et al. (2019), who highlighted that intratumoral choline levels may overlap between gliomas and metastases; however, the absence of peritumoral Cho elevation is a distinguishing feature of metastases. Although peritumoral spectroscopy was not separately analysed in our study, the relatively lower Cho/NAA values compared to gliomas support their observations.[13]

Similarly, Tsougos I et al. (2012) reported higher Cho/NAA in glioblastoma compared to metastases (3.5 vs 2.4), findings that closely parallel our results (3.15 vs 2.23). This difference can be attributed to the infiltrative nature of gliomas, in contrast to the well-circumscribed growth pattern of metastases.[14]

Meningioma

Meningiomas demonstrated the highest Cho/NAA of 5.04 ± 1.18 and Cho/Cr of 3.63 ± 0.59, along with markedly reduced NAA/Cr (0.21 ± 0.09) with near absence of NAA, consistent with their extra-axial origin and lack of neuronal tissue.

These findings are in agreement with Kousi E et al. (2012) and Huang RY et al. (2019), who reported elevated choline and absent neuronal markers in meningiomas. The high Cho/NAA values observed in our study are primarily due to the near absence of NAA, which should be interpreted in the context of tumour location.[15,16]

Medulloblastoma

Medulloblastomas in our cohort showed elevated Cho/NAA (3.48 ± 0.92), high Cho/Cr (3.05 ± 0.81), reduced NAA/Cr (0.39 ± 0.14), and prominent lipid peaks. These findings indicate aggressive tumour biology and high cellular proliferation.

These findings closely parallel Panigrahy A et al. (2013) showing significantly elevated Cho/NAA ratios in medulloblastomas compared to other posterior fossa tumours.[17] Morana G et al. (2015) demonstrated that higher Cho/Cr ratios are associated with aggressive molecular subgroups, supporting the prognostic relevance of elevated choline in our study.[18] Davies NP et al. (2020) further confirmed that high choline and lipid peaks are hallmarks of aggressive paediatric tumours, consistent with our observations.[19]

Pituitary Adenomas and Other Tumours

Pituitary adenomas in our study demonstrated moderate elevation of choline with relatively preserved NAA compared to intra-axial tumours. This reflects their non-neuronal origin and lower metabolic aggressiveness.

Stadnik TW et al. (2014) described elevated choline and absence or reduction of NAA in pituitary adenomas, findings comparable to our study[20]. Liu Z et al. (2018) reported higher Cho/Cr ratios in invasive adenomas, which aligns with the moderate choline elevation observed in our cohort[21]. Martinez Luque E et al. (2024) emphasized the importance of voxel placement in sellar lesions, which may influence spectroscopic accuracy, a factor that must be considered while interpreting results.[2]

Lipid–Lactate Peak Analysis

Lipid–lactate peaks were predominantly observed in high-grade gliomas and metastatic tumours, strongly correlating with tumour aggressiveness, with a statistically significant association with tumour grade (p < 0.001). These peaks are indicative of necrosis and anaerobic metabolism.

This is supported by Shakir TM et al. (2022), who highlight lipid peaks as markers of high-grade pathology.[10]

Correlation with Histopathology

MRS demonstrated strong correlation with histopathology, with:

• Sensitivity: 91.2%
• Specificity: 79.4%
• Diagnostic accuracy: 85.3%

These findings are comparable to Law et al. (2003) and Kim et al. (2006), validating the role of metabolite ratios in tumour grading. The high sensitivity indicates excellent ability of MRS to detect high-grade tumours, while moderate specificity reflects some overlap between tumour types.[11,12]

Strengths of the Study

Firstly, it included histopathological confirmation in all cases, ensuring a reliable gold standard for correlation.

Secondly, a comprehensive evaluation of multiple tumour types was performed, allowing comparison across a broad spectrum of brain tumours.

Thirdly, both quantitative metabolite ratios and qualitative lipid–lactate assessment were analysed, providing a holistic evaluation of tumour metabolism.

Additionally, the study demonstrated high diagnostic sensitivity and good overall accuracy, reinforcing the clinical applicability of MRS. The use of standardized spectroscopic parameters and consistent methodology further enhances the reproducibility of the findings.

Limitations

• The sample size was relatively small, and some tumour subtypes had limited representation.
• The use of a 1.5T MRI system may have limited spectral resolution compared to higher field strengths.
• Technical factors such as Partial volume effects voxel placement, voxel contamination, and spatial mismatch between MRS and histopathology may have influenced results, susceptibility artifacts may also influence spectral quality.
• Peritumoral metabolic analysis was not performed, which could have improved differentiation between gliomas and metastases.

Overall Interpretation

The present study demonstrates strong concordance with existing literature demonstrating that MRS provides reliable metabolic differentiation of brain tumours. Elevated Cho/NAA and Cho/Cr ratios correlate with tumour aggressiveness, reduced NAA/Cr reflects neuronal destruction, and lipid–lactate peaks indicate necrosis and high-grade pathology.

The statistically significant differences observed across tumour groups (p < 0.001) and strong correlation with histopathology highlight the value of MRS as adjunct to conventional MRI and reinforces its role y as a reliable non-invasive tool for tumour characterization, differentiation, and prognostic assessment.

The present study demonstrates that Magnetic Resonance Spectroscopy (MRS) provides valuable metabolic information that complements conventional MRI in the evaluation of brain tumours. Distinct metabolite patterns were observed across tumour types, with high-grade tumours showing elevated Cho/NAA and Cho/Cr ratios along with reduced NAA/Cr, reflecting increased cellular proliferation and neuronal loss. In contrast, low-grade tumours exhibited relatively lower choline ratios and preserved NAA levels.

The presence of lipid–lactate peaks was predominantly associated with aggressive tumours and areas of necrosis, suggesting their potential role as indicators of tumour aggressiveness and poorer prognosis. Therefore, the incorporation of lipid–lactate peaks further improved the identification of high-grade lesions, highlighting the importance of a multiparametric approach. A clear relationship was observed between tumour grade and spectroscopic metabolite patterns, with increasing tumour grade associated with higher choline-based ratios and progressive reduction of neuronal metabolites.

A strong concordance was observed between spectroscopic findings and histopathological diagnosis, supporting the reliability of spectroscopic metabolic markers in tumour characterization and grading.

However, overlap in metabolite ratios between tumour types and technical limitations necessitate cautious interpretation. Therefore, MRS should be used as an adjunct rather than a standalone diagnostic tool.

Overall, the study highlights that Magnetic Resonance Spectroscopy significantly enhances the diagnostic capability of conventional MRI by providing non-invasive insight into tumour metabolism. It facilitates improved tumour differentiation, assists in grading, and provides important prognostic information that may guide surgical planning and therapeutic decision-making. Integration of MRS into routine neuroimaging protocols can therefore improve diagnostic confidence and contribute to more accurate non-invasive tumour assessment. Further studies with larger sample sizes and advanced techniques are required to standardize metabolite thresholds and expand its clinical applicability.

1. Johnson DR, et al. Part I—Key concepts and the spectrum of diffuse gliomas. Radiology. 2022.
2. Martinez Luque E, et al. An update on MR spectroscopy in cancer management. Radiol Cancer. 2024.
3. Weinberg BD, et al. Clinical applications of magnetic resonance spectroscopy in neuro-oncology: review. Radiology. 2021.
4. Li X, et al. Super-resolution whole-brain 3D MR spectroscopic imaging to map D-2HG and tumor metabolism. Radiology. 2020.
5. Huang RY, Lin A. Whole-brain MR spectroscopy imaging of brain tumor metabolites. Radiology. 2020;294(3):598–599.
6. Katsura M, et al. Recognizing radiation-induced changes in the central nervous system. Radiographics. 2021.
7. Liserre R, et al. MR spectroscopy in pediatric neuroradiology: applications and pitfalls. Radiographics. 2021.
8. Galijasevic M, et al. Magnetic resonance spectroscopy in diagnosis and follow-up of gliomas. Cancers (Basel). 2022.
9. Martucci M, et al. Magnetic resonance imaging of primary adult brain tumors. Radiology. 2023.
10. Shakir TM, et al. Role of proton MR spectroscopy in grading of gliomas. Eur J Radiol Open. 2022;9:100409.
11. Law M, Yang S, Wang H, et al. Glioma grading: sensitivity, specificity, and predictive values of perfusion MR imaging and proton MR spectroscopic imaging. Radiology. 2003;228(2):382–389.
12. Kim JH, Chang KH, Na DG, et al. Proton MR spectroscopy as a diagnostic tool in differentiating low-grade and high-grade gliomas. AJNR Am J Neuroradiol. 2006;27(5):1102–1108.
13. Claus EB, et al. High-grade gliomas and solitary metastases: differentiation using perfusion and proton spectroscopic MR imaging. Radiology. 2019;292:301–310.
14. Tsougos I, et al. Differentiation of glioblastoma multiforme from metastatic brain tumors using proton MR spectroscopy. AJNR Am J Neuroradiol. 2012;33:1432–1438.
15. Kousi E, et al. Proton MR spectroscopy of intracranial meningiomas at 3T. AJNR Am J Neuroradiol. 2012;33:195–200.
16. Huang RY, et al. Imaging advances in intracranial meningiomas. Neuro Oncol. 2019;21(Suppl 1):i44–i56.
17. Panigrahy A, Blüml S. Neuroimaging of pediatric brain tumors: MR spectroscopy and metabolic markers. Radiol Clin North Am. 2013;51(4):699–713.
18. Morana G, et al. MR spectroscopy of pediatric brain tumors: metabolite patterns and molecular subgroup correlation. Neuroradiology. 2015;57:945–956.
19. Davies NP, et al. Advances in MR spectroscopic imaging of pediatric brain tumors. NMR Biomed. 2020;33:e4230.
20. Stadnik TW, et al. Proton MR spectroscopy of sellar and parasellar lesions. AJNR Am J Neuroradiol. 2014;35:1441–1446.
21. Liu Z, et al. Metabolic profiling of pituitary adenomas using 3T MR spectroscopy. Eur Radiol. 2018;28:348–356.