Abstract

Background: Radiogenomics, the intersection of imaging and genetics, is important in improving glioma diagnosis and treatment. This study aims to correlate imaging features with isocitrate dehydrogenase (IDH) mutation status, providing a non-invasive diagnostic tool to identify the genetic background of gliomas.

Methods: In a retrospective sample of 59 patients with either IDH wild-type (WT) or IDH mutant gliomas, the study employed volumetric and morphologic magnetic resonance imaging (MRI) analyses to discern molecular alterations based on radiographic signatures. Key imaging biomarkers, such as the T2/fluid-attenuated inversion recovery mismatch, contrast enhancement patterns, and diffusion/perfusion metrics, were evaluated for their ability to differentiate between IDH-WT and mutant gliomas. Receiver operating characteristic curves were employed to evaluate diagnostic performance, and a logistic regression model was developed for patient classification based on imaging.

Results: The results revealed that IDH mutant gliomas frequently exhibited distinct imaging characteristics, such as homogenous hyperintense T2 signals and absence of contrast enhancement. In addition, perfusion and diffusion metrics varied significantly between the IDH-WT and mutant groups, offering potential radiogenomic markers. A logistic regression model was developed to predict IDH status with high accuracy, identifying factors such as tumor enhancement size, presence of central necrosis, peritumoral edema, and patient age.

Conclusion: The study’s results affirm the significance of radiogenomic correlations in predicting IDH status, resonating findings from prior research. We highlight the necessity of a multimodal approach in MRI analysis to enhance the non-invasive diagnostic accuracy for glioma patients.

Keywords: Glioma, Isocitrate dehydrogenase-mutant, Magnetic resonance imaging, Non-invasive tool, Radiogenomics

INTRODUCTION

The field of neuro-oncology has witnessed a paradigm shift in the classification of glioblastomas (GBMs), driven by remarkable advances in our understanding of their molecular biology. Historically, the classification of GBMs primarily relied on histopathological features. However, recent years have seen a transition toward a molecular-based approach, acknowledging the complexity and heterogeneity of these tumors. This new classification paradigm integrates genetic and molecular characteristics, thereby improving diagnostic accuracy and modifying treatment approaches.[ 15 , 26 ] In adult type gliomas, the distinction between IDH-wild-type (WT) and IDH-mutant (IDH-mut Grade 4 Astrocytomas) GBMs is necessary and influences therapeutic approaches and patient outcomes.[ 10 ]

The concept of radiogenomics began to emerge as the integration of radiology and genomics became increasingly important in understanding and managing cancer. Radiogenomics links the appearance of tumors on radiological images, such as magnetic resonance imaging (MRI) or computed tomography (CT) scans, with their genomic characteristics. The need for radiogenomics stemmed from the realization that genetic mutations within tumors could manifest as distinct imaging features. This correlation provides insights into tumor behavior, aggressiveness, and response to treatment, which are important in adjusting patient-specific therapeutic strategies.[ 1 , 3 ] Moreover, radiogenomics in GBM has the potential to track tumor evolution and response to therapy over time. By monitoring changes in imaging characteristics, clinicians can infer changes in the tumor’s profile, providing new concepts into the effectiveness of the treatment and the need for its modification.[ 20 ]

Some studies advanced the field of radiogenomics by using machine learning algorithms to classify GBMs based on their molecular characteristics, deciphered from multiparametric MRI data. This included analyzing MRI features such as tumor volume, edge sharpness, and internal texture to predict molecular subtypes. Their machine learning model presented a step towards automating the interpretation of complex imaging data, leading to more accurate and efficient diagnoses.[ 5 ]

In this study, we conducted a retrospective analysis of patients with IDH-mutant, IDH-mutant with 1p/19q co-deletion, and IDH-WT gliomas. We aimed to correlate the histological diagnosis with imaging-based diagnosis (radiogenomics), seeking to understand the relationship between molecular alterations and their radiographic signatures. The findings have the potential to improve the precision of non-invasive diagnostic techniques, contributing to the personalized management of glioma patients.

MATERIALS AND METHODS

We conducted a retrospective cohort study at a tertiary University Hospital, focusing on patients who underwent surgical resection of gliomas in the year of 2018 and 2019. The study included individuals aged 18 and over. We excluded cases lacking incomplete MR imaging, missing pre-resection MRI, tumors located infratentorial, patients with a history of reoperation, and those previously treated with chemotherapy or radiotherapy. All tumors were reclassified according to the latest World Health Organization 2021 classification of central nervous system tumors, differentiating them into IDH-WT, IDH mutant categories, and excluding others. Furthermore, the IDH mutant gliomas were subdivided into two distinct groups: IDH mutant with or without 1p/19q codeletion.

MRI scans were performed using a uniform protocol on a 1.5 and 3.0 Tesla MRI system. All patients had a preoperative MRI at our hospital. Standardized sequences included volumetric acquisitions using a T1-weighted Magnetization Prepared Rapid Gradient Echo sequence. This volumetric data were subsequently reconstructed into three anatomical planes – axial, sagittal, and coronal. In addition, T2-weighted, T2*-weighted, and fluid-attenuated inversion recovery (FLAIR) sequences were performed with a slice thickness of 5 mm and an interslice gap of 1 mm. Diffusion-weighted imaging (DWI) was conducted with b values of 0, 500, and 1000 s/mm², and perfusion-weighted imaging (PWI) utilized dynamic susceptibility contrast-enhanced sequences.

We collected data from patient demographics, such as age and sex, and we analyzed imaging characteristics from preresection MRIs, capturing imaging appearances of the tumor on T1, T2, and T2* sequences, and the presence of T2/FLAIR mismatch (T2/FLAIR mismatch was defined as a lesion with >50% of its volume demonstrating homogeneous high signal on T2-weighted images and relatively hypointense signal on FLAIR except for a hyperintense peripheral rim).[ 7 , 20 ] The evaluation of contrast uptake on T1-weighted images included several patterns: peripheral enhancement, which shows a ring-like uptake indicative of proliferating tumor cells at the periphery of a necrotic core; nodular enhancement, presenting as well-defined, rounded area; homogeneous enhancement, where the contrast is evenly distributed throughout the lesion; heterogeneous enhancement, a mix of high and low contrast uptake areas; and, finally, the absence of enhancement. We proceeded to the identification of the affected cerebral hemisphere and the specific lobar location (e.g., frontal and temporal) of the tumor and its potential multifocality (as defined by multiple discrete areas of contrast-enhancing tumor, either embedded within or connected by T2/FLAIR signal abnormalities).[ 18 , 21 ] The radiological examination was collected from the neuroradiology reports and subsequently reviewed by an experienced neuroradiologist, utilizing the radiology department’s Picture Archiving and Communication System. In cases of initial discrepancy or doubt, a second neuroradiologist was available for consultation. We then calculated the volumetry of peritumoral edema, the size of the tumor, and the volume of contrast uptake (if any). We determined the ratio of peritumoral edema to tumor volume, termed as Peritumoral Edema Relative to the size of the tumor (PERT), to assess the potential correlation between edema and tumor size with IDH mutation statu (PERT [%] = [Peritumoral edema volume ÷ Tumor volume] × 100). Edema was classified as minimal or absent when it constitutes less than 15% of the tumor’s total volume. Intermediate edema is defined as edema that accounts for between 15% and 80% of the total tumor volume in line with prognostic thresholds described by Wu et al.[ 25 ] Finally, high/increased edema is categorized as edema that exceeds 80% of the total volume of the tumor. This segmentation allows for a detailed and quantitative assessment of the edema surrounding GBMs. Volumetric measurements were conducted using Brainlab Elements SmartBrush (Brainlab AG, Munich, Germany) which provided semi-automated tools to delineate the tumor and peritumoral edema boundaries. The software enabled the calculation of volumes through the segmentation process, which involved manual initialization followed by automated contouring. This method allowed for the accurate determination of tumor size, edema volume, and the volume of contrast uptake, subsequently used to calculate the PERT ratio [ Figures 1 and 2 ].

Figure 1:

Volumetric analysis of tumor and contrast enhancement in glioma MRI scans depicting volumetric segmentation of a glioma using Brainlab Elements SmartBrush (Brainlab AG, Munich, Germany). The total tumor volume (TV) is outlined in yellow, while the cystic component (CC), representing non-enhancing regions, is demarcated in orange. The volume of contrast enhancement is calculated by subtracting the CC from the TV (Enhancement: TV-CC). This approach allows for precise measurement of the tumor and its enhancing features, facilitating a detailed radiological evaluation.

Figure 2:

Peritumoral edema volumetry in glioma magnetic resonance imaging (MRI) coronal MRI T2 slice illustrating the segmentation of peritumoral edema in red, as analyzed using Brainlab Elements SmartBrush (Brainlab AG, Munich, Germany). The highlighted area indicates the calculated volume of the edema surrounding the glioma.

Finally, focusing on diffusion and perfusion in GBMs, we have chosen the apparent diffusion coefficient (ADC) and relative cerebral blood volume (rCBV) as our primary imaging biomarkers. On the other hand, rCBV, originated from PWI, measures the volume of blood within an area of the brain.[ 12 , 17 , 24 ] We began by reviewing the neuroradiological reports that employed a standardized approach to quantify the ADC and rCBV. The normalization of ADC and rCBV values against normal-appearing white matter is a well-established practice in our institution. It is aimed at reducing inter-patient and inter-scanner variability, ensuring that our measurements are consistent and comparable. Following this initial review, we categorize the ADC and rCBV values into three levels: low, intermediate, and high. Despite the technique being widely adopted, it is recognized that the normalization process is inherently subjective, which may contribute to variability in the interpretation of the data.[ 2 , 13 ] We sought confirmation from an additional neuroradiologist to ensure the accuracy and reliability of this specific findings.

Statistical analyses were performed with the Statistical Package for the Social Sciences, version 27.0. For descriptive statistics, medians and quartiles or frequencies and percentages were presented accordingly. To assess associations between categorical variables, such as MRI characteristics and tumor types (IDH-mutated vs. IDH-WT), a Chi-square test was used. Receiver operating characteristic curves were implemented to assess sensitivity, specificity, and precision. A logistic model was created attempting to create a tool to classify patients according to image results. The linear predictor of this model was used as continuous variable, and a cut-off was established to classify the patients according to the Youden index.

RESULTS

In this study, we recruited a cohort of 59 patients who were evaluated for glioma subtypes based on IDH mutation status. The group comprised 39 individuals with IDH-WT gliomas and 20 with IDH-mutant gliomas. Among the IDH-mutant category, seven patients were diagnosed with astrocytoma IDH-mutant, while 13 were identified as having oligodendroglioma with the IDH mutation and 1p/19q codeletion [ Table 1 ].

Table 1:

Radiographic and demographic characteristics of brain tumor patients by IDH mutation status.

Our analysis highlighted notable differences in demographic and radiographic features when categorizing patients according to their IDH mutation status. The median age across the entire cohort was 61 years, with a marked age difference observed between the IDH-WT and IDH-mutated groups. Patients with IDH-WT tumors had a higher median age of 64 years compared to 46.5 years for those with IDH-mutated tumors (P < 0.001). There was no significant difference in sex distribution between the two groups (P = 0.758), with males comprising 62.7% of the cohort.

MRI imaging characteristics presented notable differences. On T2-weighted images, 49.2% of the total cohort had hyperintense mainly homogeneous signals, with IDH-mutated tumors more frequently exhibiting this characteristic (70.0%) compared to IDH-WT (38.5%), which was statistically significant (P = 0.022), as determined by a Chi-square test. In the assessment of T1-weighted images with contrast, a clear disparity was observed in the absence of contrast enhancement, which was noted in 70.0% of the IDH-mutated but only in 2.6% of the IDH-WT tumors (P < 0.001).

T2/FLAIR mismatch was exclusively present in the IDH-mutated group, with 30.0% exhibiting this feature and none in the IDH-WT group, suggesting a strong correlation with the mutation status (P < 0.001). The location of tumors provided additional discriminative data; frontal lobe tumors were predominantly IDH-mutated (95.0%), whereas temporal, parietal, and other locations were more commonly associated with IDH-WT (P < 0.001).

Looking at perfusion metrics, a vast majority of the IDHWT tumors show high perfusion on rCBV imaging (87.5%), while the IDH-mutated tumors tend to exhibit low perfusion (60%), with the statistical significance marked at P < 0.001. The diffusion metrics on ADC/DWI imaging also exhibit considerable differences, with a high ADC value in 85% of the IDH-mutated group, frankly, contrasting with 7.7% in the IDH-WT group, indicative of a potential radiogenomic correlation (P = 0.015).

The evaluation of peritumoral edema relative to tumor size indicates that a substantial majority of IDH-mutated tumors (75%) present with minimal or no edema, whereas more than half of the IDH-WT tumors (51.3%) are associated with high peritumoral edema (P < 0.001).

Table 2 underscores the diagnostic potential of MRI features in differentiating between IDH-WT and IDH-mutated brain tumors. The analysis showed that tumor location has a significant predictive value, with frontal lobe tumors overwhelmingly being IDH-mutated (95.0%), as opposed to tumors in temporal, parietal, or other locations. This localization yielded a high sensitivity of 84.6% and an exceptional specificity of 99.5%, corresponding to an area under the curve (AUC) of 89.8%, underscoring the strong diagnostic performance for IDH mutation status.

Table 2:

Diagnostic performance of MRI features in predicting IDH mutation status.

The contrast-enhanced T1 MRI sequences also demonstrated notable discriminative power. The absence of contrast enhancement was predominantly seen in IDH-mutated tumors (70.0%), translating into a high sensitivity of 97.4% and a specificity of 70.0% for identifying IDH mutations, with an AUC of 83.7%. Conversely, heterogeneous, or other types of contrast enhancement were more common in IDH-WT tumors.

Regarding the size of the enhancing tumor, a threshold of 0 cm³ discriminated between the groups with an 89.7% sensitivity for detecting IDH-WT tumors when the enhancement size was equal to 0 cm³, and a 75.0% specificity for identifying IDH-mutated tumors when the enhancement size was 0 cm³, leading to an AUC of 82.4%. The presence of central necrosis also played a critical role, with an absence of necrosis being indicative of IDH-mutated status (sensitivity of 89.7%, specificity of 80.0%, and AUC of 84.9%). We further highlighted FLAIR mismatch as a feature, which, while not present in IDH-WT tumors, appeared in 30.0% of IDH-mutated tumors, offering a 100.0% specificity for IDH-WT and a sensitivity of 30.0% for IDH mutations. The AUC for FLAIR mismatch stood at 65.0%, indicating a moderate discriminative ability. Perfusion imaging exhibited a strong correlation with IDH mutation status. Low perfusion was significantly associated with IDH-mutated tumors, showing a sensitivity of 96.9% and a specificity of 66.6%, which corresponded to an AUC of 76.2%. Of the patients, 10.3% had no minimal or no detectable peritumoral edema (defined as <15% of peritumoral edema compared to the relative tumor size), with a larger proportion of these cases (75.0%) corresponding to IDH-mutated tumors. This was associated with a sensitivity of 89.7% and a specificity of 25.0% for predicting the presence of an IDH mutation, revealing an AUC of 82.4%. Conversely, the presence of at least ≥15% relative to the size of the tumor was observed in 89.7% of the cohort and was more common in IDH-WT tumors, accounting for 25.0% of cases. Age was another critical factor; patients younger than 54.5 years were more likely to have IDH-mutated tumors, reflected in a sensitivity of 92.3% and a specificity of 75.0% for predicting IDH mutation status, with an AUC of 83.7%.

We sought to delineate and compare the radiological alterations across the two distinct types of gliomas, IDH mutant with 1p/19q codeletion (codel), and IDH mutant without 1p/19q codeletion (IDH-Mut/-). The comparative analysis revealed variations in MRI signal characteristics, perfusion and diffusion measures, and other key radiographic features, which are detailed in Tables 3 and 4 . These tables summarize the distinctions of each tumor type’s radiographic signature, offering a comprehensive overview of the imaging biomarkers that correlate with the IDH mutation status.

Table 3:

Comparative radiographic profiles of IDH-mutated with or without codeletion versus IDH-WT gliomas.

Table 4:

Comparative radiographic features of IDH-mutated with or without codeletion versus IDH-WT gliomas.

A logistic regression model, incorporating the size of enhancement tumor (>0), size of central necrosis (>0), >15% of peritumoral edema relative to the size of the tumor, and age (>54.5 years), effectively predicted IDH-WT status. The model showed a high sensitivity of 95.0% and a specificity of 89.7% and precision 92.4% for predicting IDH-WT status when the linear predictor value was <−1,177, achieving an AUC of 92.4%, indicating excellent predictive accuracy.

DISCUSSION

Our research has revealed insights into the MRI imaging characteristics that correlate with IDH mutation status in gliomas, which are integral to the stratification of tumor subtypes due to their prognostic significance. The core of our findings rests on the evident imaging biomarkers, particularly the hyperintense signals on T2-weighted MRI scans and the absence of contrast enhancement in T1-weighted images among IDH-mutant gliomas. The absence of contrast enhancement in T1-weighted images presented with a sensitivity of 97.4% and a specificity of 70.0% for identifying IDH mutations, highlighting its potential as a robust diagnostic indicator. Similarly, the frontal lobe tumor location, which was predominantly associated with IDH mutations, had a sensitivity of 84.6% and an exceptional specificity of 99.5%, reflecting its strong diagnostic proficiency. IDH-mutated tumors were significantly more likely to exhibit homogeneous hyperintense signals on T2-weighted images, a feature present in 70% of these tumors compared to only 38.5% of IDH-WT tumors. The hyperintense T2-weighted signals in IDH-mutant tumors are especially noteworthy, as they align with the known pathophysiological alterations induced by IDH mutations.[ 14 , 22 ] These mutations lead to the accumulation of 2-hydroxyglutarate, which contributes to widespread metabolic disturbances, including aberrations in cellularity and matrix remodeling, potentially explaining the hyperintensity observed on T2-weighted MRIs. The development of a routine spectroscopic peak of 2-hydroxyglutarate could be an important biomarker. Alterations in the tumor microenvironment may hinder the accumulation of contrast agents, offering a plausible explanation for the reduced enhancement seen in T1-weighted imaging of IDH-mutant gliomas.[ 4 , 22 ]

Kickingereder et al. focused on MRI radiomic features to anticipate the status of IDH1 mutation.[ 11 ] By analyzing the MRI scans, they derived a pattern of imaging traits that correlated with the presence of IDH1 mutations. This study highlighted the potential for specific MRI features, such as contrast enhancement patterns, signal intensity, and the distribution of tumor cells, to act as surrogate markers for genetic alterations.[ 11 ] Other studies focused on T2/FLAIR mismatch sign as an indicator for the presence of IDH mutations in gliomas lacking 1p/19q co-deletions. Data from a comprehensive review and meta-analysis reveal that this sign’s sensitivity is relatively low at 42%, yet its specificity reaches 100%, which means it is a highly definitive indicator of this subtype when it does occur. The consistency in identifying the T2-FLAIR mismatch across various observers is strong, with a correlation coefficient of 0.87, reflecting a near-unanimous consensus.[ 4 , 19 ]

The T2/FLAIR mismatch was exclusively present in the IDH-mutated gliomas within our study cohort, which was observed in 30% of the IDH-mutated group. Notably, none of the IDH-WT gliomas exhibited this mismatch. The sensitivity of this feature for IDH mutations in our study was 30.0%, with a specificity of 100.0%, indicating that while it may not be present in all IDH-mutated gliomas, its presence strongly predicts the IDH mutation status. This mismatch, characterized by a hyperintense signal on T2-weighted images and a relative hypointensity on FLAIR sequences, may reflect a unique interplay between tumor physiology and IDH mutation-driven oncogenic processes, marking it as an imaging hallmark for IDH mutation status.[ 9 ]

In our cohort, IDH-WT gliomas predominantly displayed high perfusion on rCBV imaging (87.5%), while IDH-mutant gliomas typically showed low perfusion (60%). The contrast in perfusion patterns between the two groups (P < 0.001) suggests that IDH mutations may be associated with angiogenic profiles and vascular architecture that differ from those in IDH-WT gliomas. This finding correlates with the literature suggesting that IDH-mutant gliomas tend to be less angiogenic and exhibit lower rCBV due to less vascular proliferation, which is a hallmark of less aggressive tumor biology.[ 11 ]

Diffusion metrics, specifically high ADC values, were more frequently observed in the IDH-mutant group (85%) compared to the IDH-WT group (7.7%). The increase in cellular density is a distinctive marker of high-grade gliomas, in contrast to low-grade gliomas. The increase in cellular density results in a restriction of diffusibility, contrary to the facilitation of such diffusibility. This observation is fundamental for understanding the dynamics of IDH-mutant gliomas. IDH-mutant gliomas exhibit greater restriction to the diffusion of water molecules, a characteristic attributable to their cellular composition. These gliomas are composed of cells with lower proliferative capacity, biologically and metabolically more stable, and less prone to necrotic transformation. The absence or reduction of necrotic areas, which typically facilitate the diffusibility of water molecules, reinforces this hypothesis. Therefore, the presence of necrotic areas in high-grade tumors increases diffusibility due to cellular disintegration and the formation of extracellular spaces. In contrast, low-grade tumors, especially those with IDH mutations, maintain a more intact cellular structure and controlled proliferation, resulting in greater diffusion restriction.[ 6 ]

The presence of central necrosis was indicative of IDH-WT status. Specifically, the absence of necrosis showed a high sensitivity of 89.7% and a specificity of 80.0% for predicting IDH mutation. These metrics are significant as they suggest that more tumors without central necrosis are likely to be IDH-mutant. This is consistent with previous research which has indicated that IDH-mutant gliomas typically exhibit less aggressive features and less necrotic tissue compared to IDH-WT gliomas, which are more likely to display necrotic changes due to their higher grade and more aggressive nature.[ 26 ]

The study results demonstrated that a substantial majority of IDH-mutated tumors (75%) presented with minimal or no peritumoral edema, whereas more than half of the IDHWT tumors (51.3%) were associated with high peritumoral edema. This distinction aligns with the understanding that IDH-WT gliomas, often more aggressive and higher-grade, are associated with more extensive vasogenic edema due to blood-brain barrier disruption.[ 23 ] The correlation between peritumoral edema and tumor biology has been explored in literature, underscoring that peritumoral brain edema may reflect the biological aggressiveness of the tumor.[ 25 ]

The size of the enhancing tumor was also a discriminating factor; a threshold of 0 cm³ effectively distinguished between IDH-mutant (no contrast enhancement) and WT groups. The sensitivity and specificity for detecting IDH-WT tumors when the enhancement size was ≥0 cm³ were high, leading to an AUC of 82.4%. This suggests that larger enhancing tumors are more likely to be IDH-WT, a finding that supports literature indicating that IDH mutations are more common in smaller, lower-grade gliomas, while larger tumors with more significant enhancement patterns are characteristic of higher-grade, IDH-WT gliomas.[ 16 ]

The predictive model developed in this study is grounded in the need for precise, non-invasive prognostication tools in the management of glioma patients. The primary aim was to predict IDH mutation status – a key factor influencing therapy and prognosis – using preoperative MRI features. We initiated model development by selecting variables that previous studies and biological plausibility indicated as significant. Each of these factors has been associated with IDH mutation status and, therefore, was hypothesized to contribute predictive power to the model. The strength of the model lies in its high sensitivity, specificity, and predictive value. Hence, a patient with no tumor enhancement, without central necrosis, less than 15% of peritumoral edema relative to the size of the tumor and age <54.5 years effectively predicted IDH mutant status with a precision of 92,4%.

The study presented here, while offering valuable insights into the predictive characteristics of IDH mutation status in glioma patients, is subject to several limitations inherent to its design. First, the retrospective nature of the study introduces a selection bias, as it is not representative of a general population but rather a snapshot of a specific patient cohort within a given time frame. Such criteria, while refining the study group, exclude a subset of patients that could provide additional insights into the variability of imaging features post-treatment. With a relatively small sample size of 59 patients, there is a concern for potential underpowering of the study, which may affect the statistical significance and the strength of the conclusions drawn. Nonetheless, statistical significant differences were observed. Subjectivity in image interpretation, despite standardization and quantitative methodologies, remains a challenge. Variability among observers in assessing imaging characteristics can introduce inconsistencies, potentially affecting the reliability of the conclusions. The use of normal-appearing white matter as a reference in the assessment of ADC and rCBV values could introduce additional variability. Intratumoral heterogeneity poses another significant challenge, particularly in the accurate segmentation of tumor boundaries and associated peritumoral edema. Finally, this study did not address clinical endpoints such as patient survival or response to therapeutic interventions.[ 8 ] The focus was solely on predicting IDH mutation status through imaging features.

Although this study may not provide new information, it has successfully compiled the imaging characteristics of IDH-mutated tumors in a clear and comprehensible manner. It has incorporated new volumetric analysis into the evaluation of this patient population, and while only the percentage of peritumoral edema was statistically significant in differentiation from IDH-WT tumors, this represents a noteworthy step forward. In addition, the study has emphasized the importance of a multimodal approach in the interpretation of magnetic resonance imaging, underlining the potential of combining various imaging techniques to improve diagnostic accuracy and patient care in neurooncology.

CONCLUSION

Differentiation between IDH-WT and IDH-mutant gliomas is enhanced through the use of imaging biomarkers integrated within predictive models. Imaging characteristics, such as tumor enhancement, central necrosis, and peritumoral edema, show significant sensitivity, though specificity is enhanced through the application of logistic regression. In the context of IDH-mutant astrocytomas, the diagnostic yield of the T2/FLAIR mismatch sign is limited in our study but can be reinforced by incorporating additional imaging features, providing a more diagnostic basis for accurate tumor classification.

Ethical approval:

The research/study approved by the Institutional Review Board at Centro Hospitalar Universitário São João (CHUSJ), Department of Neurosurgery, São João University Hospital, number 260/2024, dated November 21, 2024.

Declaration of patient consent:

Patient’s consent was not required as there are no patients in this study.

Financial support and sponsorship:

Nil.

Conflicts of interest:

There are no conflicts of interest.

Use of artificial intelligence (AI)-assisted technology for manuscript preparation:

The authors confirm that there was no use of artificial intelligence (AI)-assisted technology for assisting in the writing or editing of the manuscript and no images were manipulated using AI.

Disclaimer

The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Journal or its management. The information contained in this article should not be considered to be medical advice; patients should consult their own physicians for advice as to their specific medical needs.

References

1. Ak M, Toll SA, Hein KZ, Colen RR, Khatua S. Evolving role and translation of radiomics and radiogenomics in adult and pediatric neuro-oncology. AJNR Am J Neuroradiol. 2022. 43: 792-801

2. Bedekar D, Jensen T, Schmainda KM. Standardization of relative cerebral blood volume (rCBV) image maps for ease of both inter-and intrapatient comparisons. Magn Reson Med. 2010. 64: 907-13

3. Beig N, Bera K, Tiwari P. Introduction to radiomics and radiogenomics in neuro-oncology: Implications and challenges. Neuro Oncol Adv. 2021. 2: iv3-14

4. Cohen AL, Holmen SL, Colman H. IDH1 and IDH2 mutations in gliomas. Curr Neurol Neurosci Rep. 2013. 13: 345

5. Dono A, Wang E, Lopez-Rivera V, Ramesh AV, Tandon N, Ballester LY. Molecular characteristics and clinical features of multifocal glioblastoma. J Neurooncol. 2020. 148: 389-97

6. Ellingson BM, Kim E, Woodworth DC, Marques H, Boxerman JL, Safriel Y. Diffusion MRI quality control and functional diffusion map results in ACRIN 6677/RTOG 0625: A multicenter, randomized, phase II trial of bevacizumab and chemotherapy in recurrent glioblastoma. Int J Oncol. 2015. 46: 1883-92

7. Georgescu MM. Multi-platform classification of IDH-Wild-type glioblastoma based on ERK/MAPK pathway: Diagnostic, prognostic and therapeutic implications. Cancers (Basel). 2021. 13: 4532

8. Gonçalves JM, Ferreira F, Carvalho B, Polónia P, Linhares P. Survival determinants in glioblastoma: An insight into biopsy-only patient outcomes. Biomedicines. 2024. 12: 2327

9. Kapsalaki EZ, Brotis AG, Tsikrika A, Tzerefos C, Paschalis T, Dardiotis E. The role of the T2-FLAIR mismatch sign as an imaging marker of idh status in a mixed population of low-and high-grade gliomas. Brain Sci. 2020. 10: 874

10. Kickingereder P, Bonekamp D, Nowosielski M, Kratz A, Sill M, Burth S. Radiogenomics of glioblastoma: Machine learning-based classification of molecular characteristics by using multiparametric and multiregional MR imaging features. Radiology. 2016. 281: 907-18

11. Kickingereder P, Sahm F, Radbruch A, Wick W, Heiland S, Von Deimling A. IDH mutation status is associated with a distinct hypoxia/angiogenesis transcriptome signature which is non-invasively predictable with rCBV imaging in human glioma. Sci Rep. 2015. 5: 16238

12. Kim BS, Kim ST, Kim JH, Seol HJ, Nam DH, Shin HJ. Apparent diffusion coefficient as a predictive biomarker for survival in patients with treatment-naive glioblastoma using quantitative multiparametric magnetic resonance profiling. World Neurosurg. 2019. 122: e812-20

13. Korfiatis P, Erickson B. The basics of diffusion and perfusion imaging in brain tumors. Appl Radiol. 2014. 43: 22-9

14. Louis DN, Perry A, Reifenberger G, Von Deimling A, FigarellaBranger D, Cavenee WK. The 2016 World Health Organization Classification of tumors of the central nervous system: A summary. Acta Neuropathol. 2016. 131: 803-20

15. Louis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D. The 2021 WHO classification of tumors of the central nervous system: A summary. Neuro Oncol. 2021. 23: 1231-51

16. Mellinghoff IK, Chang SM, Jaeckle KA, Van den Bent M. Isocitrate dehydrogenase mutant grade II and III glial neoplasms. Hematol Oncol Clin North Am. 2022. 36: 95-111

17. Olson MP, Nair D, LaFontaine M, Chunduru P, Villanueva-Meyer JE, Cha S. Recurrent tumor and treatment-induced effects have different MR signatures in contrast enhancing and non-enhancing lesions of high-grade gliomas. Neuro Oncol. 2020. 22: 1516-26

18. Park SI, Suh CH, Guenette JP, Huang RY, Kim HS. The T2-FLAIR mismatch sign as a predictor of IDH-mutant, 1p/19qnoncodeleted lower-grade gliomas: A systematic review and diagnostic meta-analysis. Eur Radiol. 2021. 31: 5289-99

19. Patel SH, Poisson LM, Brat DJ, Zhou Y, Cooper L, Snuderl M. T2-FLAIR mismatch, an imaging biomarker for IDH and 1p/19q status in lower-grade gliomas: A TCGA/TCIA project. Clin Cancer Res. 2017. 23: 6078-85

20. Politi LS, Levi R. Editorial comment: Radiogenomics of glioblastoma: Shifting the focus from tumor cells to immune microenvironment. Eur Radiol. 2023. 33: 207-8

21. Sakai Y, Yang C, Kihira S, Tsankova N, Khan F, Hormigo A. MRI radiomic features to predict IDH1 mutation status in gliomas: A machine learning approach using gradient tree boosting. Int J Mol Sci. 2020. 21: 8004

22. Thomas T, Thakur S, Young R. Imaging 2-hydroxyglutarate and other brain oncometabolites pertinent to critical genomic alterations in brain tumors. BJR Open. 2023. 5: 20210070

23. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 2010. 17: 98-110

24. Wan B, Wang S, Tu M, Wu B, Han P, Xu H. The diagnostic performance of perfusion MRI for differentiating glioma recurrence from pseudoprogression: A meta-analysis. Medicine (Baltimore). 2017. 96: e6333

25. Wu CX, Lin GS, Lin ZX, Zhang JD, Liu SY, Zhou CF. Peritumoral edema shown by MRI predicts poor clinical outcome in glioblastoma. World J Surg Oncol. 2015. 13: 97

26. Zhang B, Chang K, Ramkissoon S, Tanguturi S, Bi WL, Reardon DA. Multimodal MRI features predict isocitrate dehydrogenase genotype in high-grade gliomas. Neuro Oncol. 2017. 19: 109-17