RPS 716 - Novel developments in neuro-oncology

Author Block: M. Bauer, S. Mangesius, M. Wagner, J. Kerschbaumer, D. Pinggera, J. Mangesius, A. E. Grams, E. R. Gizewski, C. Birkl; Innsbruck/AT
Purpose: Brain metastases exhibit complex microstructural heterogeneity, challenging accurate tissue characterisation. This study aimed to investigate whether the MRI model Neurite Orientation Dispersion and Density Imaging (NODDI) can differentiate tissue types across entire regions in brain metastases patients.
Methods or Background: This single-centre, prospective study included nine patients with 39 confirmed brain metastases, restricted to lesions of sufficient size for clinically relevant differential diagnosis. Head MRI scans were performed at 3 T using a multi-shell diffusion protocol. NODDI parameter maps were calculated for Orientation Dispersion Index (ODI), Neurite Density Index (NDI) and Free Water Fraction (FWF) using the Watson model. Whole-tissue regions representing oedema, necrosis, neoplasm, white matter, deep grey matter and cortical grey matter were segmented automatically in 3D Slicer using the MONAI deep learning pipeline on standard sequences. Univariate analyses of pairwise tissue differences employed the Wilcoxon signed-rank and paired Student’s t-test. Multivariate analyses used Hotelling’s T2, combining all three NODDI parameters to evaluate their joint discrimination power.
Results or Findings: Univariate analyses revealed multiple statistically significant differences in NODDI parameters, with the single non-significant result for all parameters when comparing neoplasm and cortex (ODI: p = .55; NDI: p = .41; FWF: p = .17). FWF and ODI showed significant differences between pathological and normal appearing regions (p < .02), suggesting their value as markers of microstructural alteration. Multivariate analyses demonstrated significant discriminations across all tissue comparisons (p < .04), except between cortex and necrosis (p = .06).
Conclusion: NODDI parameters differentiate necrosis, oedema and neoplasm in patients with clinically relevant brain metastases. Incorporating multivariate analyses offers the combined evaluation of NODDI parameters, providing complementary information.
Limitations: The used NODDI model may underestimate complex fibre dispersions in tumours. Automatic segmentation was not validated by experts.
Funding for this study: Funding was provided by SNSF Postdoc.Mobility Fellowship PM221964 to Melanie Bauer.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by institutional review board and written informed consent was obtained from all participants.

Author Block: K. Deng, Y. Xue, R. Jiang; Fuzhou/CN
Purpose: To explore the relationship among isocitrate dehydrogenase (IDH) genotypes of glioma,the diffusion MRI features of corticospinal tract (CST) and the neurologic assessment in neuro-oncology (NANO) score.
Methods or Background: 65 patients with gliomas adjacent to CST and 32 healthy controls underwent structural and diffusion MRI.Diffusion metrics, including DTI and NODDI, were calculated along the CST.Neurologic function was assessed using NANO scale.NANO score and subscores were compared between IDH wild-type and IDH mutant.The CST features were analysed between affected and healthy side for glioma patients and between left and right side for healthy controls.The relative CST features were evaluated among the three groups.Spearman rank correlation analysis was used to evaluate the correlation between the relative CST features and NANO score.
Results or Findings: Results showed significantly higher NANO scores, indicating worse function, in IDH wild-type patients compared to IDH mutant patients. Significant differences were found in motor subscores (gait, strength, ataxia, facial strength).CST microstructure on the affected side was significantly altered in both groups, with more extensive changes in IDH wild-type gliomas.Significant differences in relative CST metrics (ICVF, MD, RD) were found between IDH genotypes.There were differences in the relative CST features between the control group and different IDH genotypes.Furthermore, AD and MD correlated positively with NANO score, while ICVF and ODI correlated negatively.
Conclusion: Neurological impairment, particularly motor function,is more severe in IDH wild-type gliomas,likely due to greater CST microstructural injury. NODDI provides sensitive microstructural detection, confirming that CST integrity is better preserved in IDH mutant gliomas.
Limitations: There was no direct correlation between diffusion-sampled data and surgical findings. Therefore, it could not be ascertained whether the glioma truly infiltrated the CST. However, the decrease in motor function may indirectly reflect the injury of CST.
Funding for this study: This work was supported by Fujian Provincial Natural ScienceFoundation of China (Grant number: 2025J01800), Fujian provincial health technology project (Grant number: 2021QNA013), the Education and Research Program for Middle-aged and Young Teachers in Fujian Province (Grant number: JAT210108) .
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Institutional Review Board of Fujian Medical University Union Hospital (2020KY0132), and written informed consent was obtained from all the participants.

Author Block: G. Pentassuglia¹, M. Martucci², M. L. Angeli², F. Lisi¹, L. Ruscelli³, A. Infante², S. Gaudino¹; ¹Rome/IT, ²Roma/IT, ³Bergamo/IT
Purpose: Brain metastases are the most common intracranial tumors in adults and are critical for oncologic staging and treatment. Although MRI is the gold standard, contrast-enhanced head CT is frequently performed during whole-body oncologic follow-up and often interpreted by general radiologists, risking underdiagnosis. Maximum Intensity Projection (MIP) is a CT post-processing technique widely used in thoracic and vascular imaging but underutilized in neuroimaging. This study evaluated whether MIP can improve CT detection of brain metastases, particularly for less-experienced readers.
Methods or Background: This retrospective single-center study included 100 oncologic patients who underwent brain MRI within 30 days of contrast-enhanced whole-body CT. MRI served as the reference standard. Two radiologists—a general radiologist (R1) and a neuroradiologist (R2)—reviewed CT scans in two sessions: standard axial CT and CT with MIP reconstructions. The number of detected lesions, lesion size, and reporting time were recorded.
Results or Findings: MRI identified significantly more metastases than CT (mean ± SD: MRI = 12.4 ± 33.6; CT without MIP: R1 = 3.5 ± 5.9, R2 = 4.9 ± 10.0; CT with MIP: R1 = 3.8 ± 6.0, R2 = 4.9 ± 9.6; p < 0.001). MIP modestly improved detection for R1, while R2’s results remained stable. R1’s performance with MIP approximated R2’s with standard CT, reducing inter-reader variability. MIP also shortened report times (−1 min for R1; −2 min for R2). In oligometastatic patients (<5 lesions), CT_MIP aligned more closely with MRI, improving reliability of findings relevant for treatment planning.
Conclusion: Although MRI remains superior, MIP-enhanced CT increases lesion detection and reduces report times, particularly for general radiologists. Incorporating MIP into routine CT protocols may improve diagnostic accuracy and efficiency in oncologic follow-up, especially in non-specialist or resource-limited settings.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: The study is retrospective.

Author Block: L. Mazu¹, S. Zeng¹, W. Zhou², S. He¹, N. Zhu¹, J. Wu¹, Z. Yang¹, H. Wei², J. Chu¹; ¹GuangZhou/CN, ²Shanghai/CN
Purpose: This study aimed to (1) assess the value of quantitative paramagnetic( χpara) and diamagnetic( χdia) susceptibility in predicting glioma grading, IDH status, and prognosis; and (2)verify their correlations with intratumoral iron and tumor-associated macrophages (TAMs).
Methods or Background: We retrospectively enrolled 81 patients with diffuse gliomas (44 men, mean age 50 years) who underwent MRI including QSM. χpara and χdia were derived using a novel APART-QSM technique. Histological analysis included iron quantification (Perls stain) and multiplex immunofluorescence for TAM markers (CD68, CD86, CD163). Diagnostic models for tumor grade and IDH status were developed. Survival analysis and Spearman correlation with pathological indicators were performed.
Results or Findings: The diagnostic models for IDH status and tumor grade achieved AUCs of 0.81 (95%CI 0.71-0.90) and 0.87 (95% CI 0.78-0.97), respectively. The results showed that χdia was an independent predictor for IDH-mutant status (p=0.016.OR=0.12), while χpara was significantly associated with high-grade glioma (p=0.01.OR=8.16).Calibration curves and decision curve analyses confirmed good predictive performance and clinical utility. Kaplan-Meier analysis showed significant overall survival differences stratified by χpara/χdia. Both χpara and χdia positively correlated with iron concentration (r=0.32, r=0.29; p<0.05). Negative correlations were observed between χpara/QSM and CD163⁺ macrophage infiltration (r=-0.27, r=-0.25; p<0.05). Additionally, there was no significant correlation among the other variables.
Conclusion: χpara and χdia serve as valuable imaging biomarkers for predicting molecular subtype, grading, and prognosis in gliomas. Their correlations with iron deposition were confirmed, supporting their biological relevance. These findings may aid in non-invasive glioma characterization and personalized treatment planning.
Limitations: Not applicable.
Funding for this study: This study was funded by National Natural Science Foundation of China (NSFC 82172015)
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This retrospective study received ethical approval of the Clinical Research and Laboratory Animal Ethics Committee,
The First Affiliated Hospital of Sun Yat-sen University (number: [2024]763).

Author Block: N. Plakhotina, D. Drozdovskaya, A. V. Smirnova, D. I. Kuplevatskaya, K. Boiko, A. Brusser; Saint Petersburg/RU
Purpose: To define the imaging features and outcomes of pediatric radiation-induced brain tumors (RIBT) using a multicenter cohort. The study aimed to apply diagnostic criteria, ensure lesion-field concordance, utilize advanced imaging (MRI, amino-acid PET), and establish realistic follow-up expectations.
Methods or Background: We reviewed 14 young patients treated between 2019 and 2025 (average age, 5.4 years). To confirm these were radiation-induced tumors, we checked: did the tumor appear where radiation was given before, was it at least 4 years after radiation, was it a different tumor type than the original disease, and was there no family cancer history. Image analysis assessed field geometry, contrast enhancement, diffusion-weighted imaging, and radiopharmaceutical uptake on amino acid PET.
Results or Findings: All tumors were high-grade gliomas located within the previous radiation fields, emerging after a median latency of 7 years. MRI showed infiltrative tumors with heterogeneous contrast enhancement. Tumors were located in the cerebral hemispheres (54%), cerebellum (33%), and brainstem (13%). Differential diagnoses included neuroleukemia relapse and other encephalopathies. Amino acid PET confirmed high radiopharmaceutical uptake in the tumor foci, aiding visualization. Outcomes were poor: 7 patients died, median overall survival was 13.8 months, and progression-free survival was 8 months. Progression was predominantly local (~90%); extraneural spread was rare (0,5%). In a separate sub-cohort receiving repeat proton irradiation (n=8), median overall survival was 19 months.
Conclusion: In children, RIBT typically manifest as high-grade gliomas within irradiated areas after a long latency. Diagnosis relies on correlating radiation therapy history with modified criteria, supported by multiparametric MRI and amino acid PET. Despite modern treatments, the prognosis remains uncertain, necessitating careful, lifelong neuro-oncological monitoring for survivors of cranial irradiation. Imaging is critical for guiding biopsy timing and therapy selection.
Limitations: N/A
Funding for this study: N/A
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: S. Namkhainyam, D. Khashbat, T. Dashjamts, A. Bat-Arildii, E. Ochirjav; Ulaanbaatar/MN
Purpose: To evaluate the survival of patients diagnosed with malignant gliomas and to identify MRI imaging features associated with prognosis.
Methods or Background: The study was conducted between 2018 and 2022 by sampling 125 clients who were diagnosed with malignant glioma brain tumor by undergoing MRI scan. Retrospective cohort analysis was done selecting 34 cases according to the inclusion and exclusion criteria.
Results or Findings: Total of 34 cases, 28 (60.5%) were male and 6 (29.5%) were female, with a median age of 50.67 ± 19.43 years. The median survival time after diagnosis was 15.7 months. For patients diagnosed with WHO Grade III malignant gliomas, the median survival time was 35.9 ± 10.1 months, while for those with Grade IV malignant gliomas, it was 12.2 ± 8.2 months. The overall survival rates among the study participants were 92% at 1month, 82% at 3months, 67% at 6months, 61% at 9months, 60% at 1 year, 12% at 3 years, and 0% at 5 years.
According to the comparison of survival rate with MRI imaging features were significantly associated with prognosis: high tumor square (P=0.011), multiple hemorrhagic components (P=0.047), high edema index (P=0.007), necrosis (P=0.0001), thickness of contrast enhancement (P=0.003) and midline shift (P=0.019).
Conclusion: The average survival time following diagnosis of high grade glioma was 15.7 months. For patients with WHO grade III gliomas, the mean survival time was 35.9 ± 10.1 months, whereas for grade IV gliomas 12.2 ± 8.2 months.
Patients with WHO grade IV gliomas had significantly shorter survival than those with grade III gliomas. MRI features such as tumor volume, peritumoral edema index, necrosis, hemorrhagic content, and mass effect were found to be negative prognostic factors for survival.
Limitations: Our study was conducted onrelatively small number ofcases.
Funding for this study: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. ID:2024/3-06
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by ethics Committee pf the Mongolian National University of Medical Sciences and National Pathology center of Ulaanbaatar Mongolia

Author Block: V. Shahmaei¹, F. Sodaei²; ¹Tehran/IR, ²Odense/DK
Purpose: Accurate preoperative grading of gliomas is crucial for treatment planning and prognosis. Conventional MRI lacks specificity, whereas quantitative susceptibility mapping (QSM) sensitively detects magnetic susceptibility changes from iron, calcification, and microhemorrhage. Intratumoral susceptibility signals (ITSS), reflecting microvascular proliferation and hemorrhage, are closely associated with tumor angiogenesis and aggressiveness, offering a promising noninvasive biomarker for glioma grading.
Methods or Background: A retrospective analysis was conducted on 40 pediatric patients with histopathologically confirmed gliomas, categorized into low-grade (WHO grade II) and high-grade (WHO grade III–IV) groups. Multi-echo T2* gradient-echo MRI data were processed using the Morphology Enabled Dipole Inversion (MEDI) QSM pipeline. Tumor regions were delineated, and both histogram-derived susceptibility metrics (mean, percentiles, skewness, kurtosis) and ITSS counts were quantified. Statistical comparisons between groups were performed using independent-samples t-tests, and diagnostic performance was assessed with receiver operating characteristic (ROC) analysis.
Results or Findings: High-grade gliomas exhibited significantly higher mean susceptibility values and greater histogram heterogeneity (90th percentile, kurtosis; p < 0.05), as well as markedly increased ITSS counts, compared with low-grade tumors. ITSS alone achieved an area under the ROC curve (AUC) of 0.87 (sensitivity 82%, specificity 83%), while histogram-derived susceptibility metrics yielded an AUC of 0.88 (sensitivity 85%, specificity 80%). The combined model, integrating both histogram features and ITSS, demonstrated the highest performance, with an AUC of 0.92, sensitivity of 90%, and specificity of 85%.
Conclusion: Histogram features derived from QSM, together with ITSS quantification, provide robust, noninvasive biomarkers for glioma grading. In this study, incorporating ITSS improved the differentiation of glioma grades, highlighting the potential of QSM as a valuable adjunct to conventional MRI for preoperative tumor evaluation.
Limitations: The limitations of the study are the retrospective design and small sample size, which may limit generalizability.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the ethics committee of Mahak Hospital.

Author Block: Y. Zhao, N. Mei, Z. Ruan, T. Xu, Y. Lu, B. Yin; Shanghai/CN
Purpose: To evaluate the feasibility of 5T SWI imaging in the diagnosis of gliomas, we conducted a comparative analysis with 3T imaging to assess image quality and determine the clinical value of 5T SWI in the preoperative evaluation of gliomas.
Methods or Background: Between 2024 and 2025, the glioma patients underwent 5T MRI scans. Each patient was individually matched with a control subject selected from the 3T MRI cohort. Radiologist scoring of image quality using a Likert scale, SNR, CNR, contrast ratio, and ITSS assessment were obtained. Paired Wilcoxon signed rank tests were used to compare characteristics derived at 3 T and 5T . The correlation between ITSS scores and molecular pathology was analyzed.
Results or Findings: A total of 27 participants were included undergoing 5T MRI. Each case was individually matched with one control subject selected from the 3.0T MRI cohort. Median image quality scores were 4.0 at 3 T and5.0 at 5T . SNR and CNR at 5 T was higher than at 3T. The degree of ITSS in 5.0 T cohort was significantly higher than that in 3T cohort . In the 5T cohort, a statistically significant association was observed between ITSS and WHO grades . The ROC curve analysis indicated that 1.5 was the best cutoff value to predict pathological grade. It shows that gliomas with a Ki-67 ≥ 10% exhibit a significantly higher ITSS than those with a Ki-67 < 10% .
Conclusion: SWI at 5T was found to be feasible and provided visualization of distal small vessel branches. And the image quality of 5T enables more precise ITSS delineation and demonstrating correlation with tumor proliferative activity
Limitations: The limitations of this study include its single-center design and small sample size, requiring future validation
Funding for this study: Not applicable
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information:

Author Block: Q. Zhou, J. Zhou; Lanzhou/CN
Purpose: Tertiary lymphoid structures (TLSs) are associated with improved survival in patients with various malignancies. However, their prognostic value and correlation with imaging features in glioblastoma (GBM) remain unclear. This study aimed to examine the relationship between TLSs and survival outcomes, and their association with imaging characteristics in patients with GBM.
Methods or Background: A retrospective analysis was conducted on 190 newly diagnosed patients with GBM. TLS-positive (TLSs+) and TLS-negative (TLSs-) groups were classified via hematoxylin-eosin staining to identify lymphocyte aggregates. Multiplex immunofluorescence staining (for CD20, CD3, CD21, and CD23) assessed TLS maturity, categorizing them into mature (mTLSs) and immature (imTLSs) subgroups. Survival was analyzed using the Kaplan-Meier method, while logistic regression developed TLS predictive models. Clinicopathological and Visually Accessible Rembrandt Images (VASARI) imaging features were compared between groups.
Results or Findings: Among 190 cases, 85 (44.73% positivity rate) were TLSs+, including 47 imTLSs and 38 mTLSs. Significant differences in overall survival were observed between TLSs+TLSs- groups and mTLSs/imTLSs groups (all p < 0.05). For VASARI features, the TLSs+/TLSs- groups differed significantly in F2, F8, F17, F22 (p < 0.05), while the mTLSs/imTLSs groups differed in F6, F16, F18 (p < 0.05); these features were independent predictors of TLS presence and maturity.
Conclusion: Conclusion: TLSs exist at different maturation stages in GBM and serve as favorable prognostic biomarkers. TLS presence and maturity are significantly correlated with patients’ overall survival and specific VASARI features. Preoperative prediction of TLSs via imaging may support risk stratification and individualized treatment for patients with GBM.
Limitations: First, this was a single-center, retrospective investigation. Second, the patient cohort size was relatively small, which may have limited the robustness of the findings.
Funding for this study: This study is supported by grants from the National Natural Science Foundation of China (grant number: 82371914), the Cuiying Science and Technology Innovation Program of Lanzhou University Second Hospital (grant number:CY2023-YB-A03),the Science and Technology Program of Lanzhou City(grant number:2024-9-122), and the Science and Technology Program of Gansu Province(grant number: 24JRRA376).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Medical Ethics Committee of our institution (Ethics Approval No. 2025A-364), and the research protocol adhered to the principles of the Declaration of Helsinki.

Author Block: J. Hadchiti¹, S. HAFFIANE¹, S. Ganjoo², M. El-AICHI¹, H. J. Howell², G. Garcia¹, N. Lassau¹, C. S. Balleyguier¹, S. Ammari¹; ¹Villejuif/FR, ²New York, NY/US
Purpose: Evaluate the predictive performance of a new set of imaging biomarkers related to microvasculature, blood flow, and tumor oxygen use in distinguishing tumor recurrence from radiation necrosis compared to the commonly used cerebral blood volume (CBV) biomarker
Methods or Background: 119 symptomatic patients with metastatic brain lesions post-radiotherapy were enrolled . 205 total lesions were analyzed across this sample. The diagnosis was tumor recurrence (n= 71 in test cohort and n = 7 in validation cohort) and radiation necrosis (n = 111 in test cohort and n = 16 in validation cohort) using expert-consensus derived from pathology and long-term follow-up as a reference standard. We studied 9 radiologist-reported imaging biomarkers (corrected CBV, non-corrected CBV, Delay, CTH, CTH Max, OEF, rLeakage, rCMRO2, COV). We performed univariate analyses to compare the area under the receiver operating characteristic curve (AUC) for the ratio values across all biomarkers in the test and validation cohorts. We additionally used a Random Forest model to identify the optimal combination of biomarkers with the highest predictive performance.
Results or Findings: In the test cohort, Delay_ratio, CTH_ratio, CTH Max_ratio , OEF_ratio , rCMRO2_ratio , and COV_ratio all significantly outperformed the predictive performance of CBV_corrected_ratio and CBV_noncorrected_ratio. The Random Forest algorithm demonstrated that the optimal model for differentiating between TR and RN was comprised of five biomarkers: CBV_corrected_ratio, CBV_noncorrected_ratio, Delay_ratio, OEF_ratio, and COV_ratio
with Delay_ratio having the highest feature importance and CBV_noncorrected_ratio having the lowest.
Conclusion: The novel imaging biomarkers investigated in this study demonstrate promising predictive performance in distinguishing between tumor recurrence and radionecrosis that warrants prospective validation
Limitations: Variation in patient numbers across different primary tumor types
Single institution study
Funding for this study: No funding
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Local ethic comimittee of Gustave Roussy Institute

Author Block: F. Yu, W. Chen, P. Chen, J. f. Chen; Chongqing, China/CN
Purpose: Methylation of the O⁶-methylguanine-DNA methyltransferase (MGMT) promoter is a key molecular marker for guiding treatment strategies and predicting prognosis in glioblastoma. This study investigated the value of diffusion tensor imaging (DTI) and arterial spin labeling (ASL) histogram parameters for non-invasively predicting MGMT status, and explored the mediating role in how MGMT status influences patient survival outcomes.
Methods or Background: We retrospectively analyzed DTI and ASL data from 349 glioblastoma patients, extracting histogram parameters including fractional anisotropy (FA), mean diffusivity (MD), and cerebral blood flow (CBF). Intergroup comparisons were performed based on MGMT status. Predictive performance for MGMT status was evaluated using ROC curves, calibration curves and the Hosmer-Lemeshow goodness-of-fit test. For survival assessment, Kaplan-Meier and Cox regression analyses evaluated associations between imaging features, molecular markers, and overall survival (OS). Bootstrap mediation analysis assessed whether imaging parameters mediated the relationship between MGMT status and survival.
Results or Findings: The DTI model (AUC = 0.842) was significantly superior to the ASL model (AUC = 0.619) in predicting MGMT status. Kaplan-Meier analysis confirmed that MGMT promoter methylation was a significant protective prognostic factor (Log-rank P = 0.041). However, Cox proportional hazards regression analysis showed that while imaging parameters were not independent OS prognostic factors, mediation analysis identified FA 10th and MD skewness as complete mediators in the association between MGMT status and OS.
Conclusion: DTI histogram parameters are effective non-invasive biomarkers for predicting MGMT status. Specifically, FA 10th and MD skewness serve as complete mediators, clarifying the key mechanism by which unmethylated MGMT drives poor prognosis through inducing microstructural disruption and increased tissue heterogeneity in tumors. This provides important imaging insights for understanding glioblastoma biology and may suggest new targets for future targeted treatment strategies.
Limitations: Not applicable
Funding for this study: Not applicable
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the Ethics Committee of the First Affiliated Hospital of the Army Medical University (KY2024109).

Author Block: A. K. Nawaz, S. Ahmed, A. Iqbal, K. Siddique; Lahore/PK
Purpose: Spinal metastasis are the most common malignant lesions of spine and major cause of neurological morbidity. Large-scale data on lesion distribution across spinal compartments and their primary tumor origins are limited. This study analyzes MRI patterns of spinal metastasis and summarizes primary tumor types in a large cohort.
Methods or Background: We retrospectively reviewed MRI reports of 12,405 patients with biopsy-proven or clinically confirmed spinal metastasis from 2020-2025. Lesions were categorized into vertebral body, paravertebral soft tissues, epidural, intradural–extramedullary and intramedullary compartments. Primary tumor types and presence of cord compression were recorded. Counts and percentages were calculated.
Results or Findings: Compartment distribution: Epidural 10,918 [88%], vertebral body 5262 [42%], paravertebral 397 [3%], intradural–extramedullary 124 [1%], intramedullary 50 [less than 1%].
Primary tumors: Lung 4962 [40%], breast 3722 [30%], prostate 1737 [14%], renal 744 [6%], thyroid 248 [2%], gastrointestinal [496 [4%], melanoma 124 [1%], sarcoma to 48 [2%], lymphoma 124 [1%].
Epidural lesions were most frequent. Rare intradural and extramedullary metastasis were observed, particularly among melanoma, renal and lung cancers.
Compartment based evaluation improved distinction from mimics and provided a structured framework for reporting in multidisciplinary planning.
Conclusion: In this large cohort, spinal metastasis demonstrate distinct MRI patterns across anatomical compartments. Systemic compartment based assessment enhances diagnostic accuracy, facilitates early detection of cord compression and informs multidisciplinary management. Rarer intramedullary lesions, though uncommon, required high vigilance.
Key points:
1. MRI is essential for comprehensive assessment of spinal metastasis.
2. Compartment analysis improves differentiation from mimics.
3. Primary tumor data provides context without implying strict compartment correlations.
4. Rarer intramedullary lesions demand careful attention.
Limitations: None
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This retrospective anonymized review was conducted in accordance with our institutional IRB guidelines. No additional patient consent was required.

Author Block: J. Liu; Henan Province/CN
Purpose: Purpose: This study was aimed to determine whether a combined diagnosis of histogram analysis of magnetic resonance imaging (MRI) based on support vector machine (SVM) can distinguish sellar/suprasellar germinomas from solid and cystic solid craniopharyngiomas.
Methods or Background: Methods: We retrospectively appraised a pathology database for 5 years and obtained 68 histopathologically proven sellar/suprasellar germinomas (n=20) or craniopharyngiomas (n=48) with magnetic resonance images. The T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), apparent diffusion coefficient (ADC) map and contrast-enhanced T1WI (CE-T1WI) of each patient were analyzed with the Mazda software. Nine parameters were selected as indicators of comparison and they were variance, skewness, kurtosis, mean, 1st percentile, 10th percentile, 50th percentile, 90th percentile and 99th percentile. The parameters with differential-diagnosis significance were used to establish the combined diagnostic model of SVM.
Results or Findings: Results: Among the nine parameters extracted using histogram analysis, the kurtosis of T1WI and the skewness, 1st percentile of T2WI and the mean, skewness, 50th percentile of CE-T1WI and almost all parameters of ADC map except variance, kurtosis and 1st percentile showed statistical significance between groups. Among the above parameters, the area under the curve (AUC) of 50th percentile of ADC map was the largest with the value of 0.916 (83.3% sensitivity and 85.0% specificity). After the analysis of combined diagnosis was performed, the AUC of SVM model in testing set showed the value of 0.969 (90.0% accuracy).
Conclusion: Conclusions: The histogram analysis of MRI may help differentiate sellar/suprasellar germinomas from solid and cystic solid craniopharyngiomas. Furthermore, the combined diagnosis of MRI histogram based on SVM was more effective than each individual histogram parameters.
Limitations: The sample size of this study is limited and needs to be expanded in future studies
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Our research received an approval from the Ethics Review Committee in Zhengzhou University, the Affiliated Tumor Hospital. Due to this research protocol requiring only a retrospective analysis of data, informed consents were waived.

Author Block: W. Jun, L. Jue, S. Gui, J. Wang; Wuhan/CN
Purpose: This study aimed to evaluate the diagnostic utility of time-dependent diffusion MRI (td-dMRI)-derived microstructural parameters in differentiating PCNSL from GBM, and to correlate these parameters with histopathologic findings.
Methods or Background: This retrospective study included 32 GBM and 19 PCNSL patients who underwent 3.0-T MRI with oscillating gradient spin-echo (OGSE) and pulsed gradient spin-echo (PGSE) sequences. Microstructural parameters (intracellular volume fraction (Vin), cell diameter, cellularity, extracellular diffusivity (Dex)) were compared between them. The area under the receiver operating characteristic curve (AUC) was used to evaluate the diagnostic performance of these indices. Histopathologic validation was performed by correlating td-dMRI parameters with hematoxylin-eosin (H&E) stained sections.
Results or Findings: In enhancing tumor regions, PCNSL showed significantly lower cell diameter and Dex, but higher Vin and cellularity than GBM (all p < 0.001). Vin demonstrated the highest diagnostic accuracy (AUC = 0.901, sensitivity = 0.737, specificity = 0.906). No significant differences were found in peritumoral regions. Vin correlated strongly with histopathologic nuclear volume fraction (r = 0.76, p < 0.001).
Conclusion: Td-dMRI-derived microstructural parameters, particularly Vin, effectively differentiate PCNSL from GBM, offering a noninvasive approach to improve preoperative diagnosis.
Limitations: First, due to the small sample size, the model may be overfitted, so combined diagnostic performance was not evaluated. Larger sample sizes are necessary to verify this model in the future. Second, ROIs were manually defined, automated tumor segmentation methods may reduce the inter-operator variability and improve the accuracy. Third, histopathologic slides were obtained by sampling a limited area of the tumor, whereas td-dMRI-based microstructural parameters were derived from the entire tumor region, leading to a lack of precise matching. Finally, we didn't combine td-dMRI parameters with other MRI methods, which may further improve the diagnostic performance of a single indicator.
Funding for this study: This work was supported by the National Natural Science Foundation of China (82202230, 82371945).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Institutional Review Board approval was obtained.