Metabolic, Perfusion, and Diffusion Imaging Enhance Diagnosis and Prognosis of H3K27-Altered Diffuse Midline Gliomas

Metabolic, Perfusion, and Diffusion Imaging Enhance Diagnosis and Prognosis of H3K27-Altered Diffuse Midline Gliomas | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Metabolic, Perfusion, and Diffusion Imaging Enhance Diagnosis and Prognosis of H3K27-Altered Diffuse Midline Gliomas Jun Qiu, Junjie Li, Yunyun Duan, Jun Sun, Yuna Li, Min Guo, Minghao Wu, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5128033/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 19 Nov, 2025 Read the published version in BMC Medicine → Version 1 posted 8 You are reading this latest preprint version Abstract Background: This study aimed to assess the contributions of metabolism, perfusion, and diffusion kurtosis imaging (DKI) to the diagnosis and prognostic prediction of H3K27 -altered diffuse midline gliomas (DMGs). Methods: Between June 2020 and May 2023, 95 patients (mean age 16.98 years; 61.1% female) with brainstem tumors, including 71 H3K27 -altered DMGs and 24 H3K27 wide-type brainstem tumors (referred as non-DMGs), underwent preoperative conventional and advanced MRI (amide proton transfer-weighted [APTw], arterial spin labeling [ASL], and DKI). Logistic and Cox regressions with leave-one-out cross-validation (LOOCV) were used to evaluate the separate and integrated contributions of advanced MRI to diagnostic and prognostic tasks. Results: Advanced MRI techniques significantly improved the diagnostic and prognostic performances for H3K27 -altered DMGs. Specifically, APTw demonstrated predictive value for both diagnosis (odds ratio [OR] = 3.81, p = 0.004) and prognosis (Hazard ratio [HR] = 1.89, p = 0.003). ASL-derived relative cerebral blood flow (rCBF) improved diagnostic (OR = 7.74, p = 0.019) and prognostic (HR = 2.87, p = 0.002) performances. DKI-derived mean diffusivity (MD) was significantly associated with the diagnosis of H3K27 -altered DMGs (OR = 0.03, p = 0.012). Integration of these metrics revealed that APTw (OR = 3.41, p = 0.014) and MD (OR = 0.05, p = 0.049) were independent diagnostic variables of H3K27 -altered DMGs, while APTw (HR = 1.66, p = 0.037) and rCBF (HR = 2.15, p = 0.038) were independent prognostic factors for overall survival (OS). Conclusion: Metabolic, perfusion, and diffusion imaging can improve the diagnosis and prognosis of H3K27 -altered DMGs beyond conventional MRI, which may aid clinical decision-making. H3K27-altered diffuse midline gliomas amide proton transfer-weighted imaging (APTw) arterial spin labeling (ASL) diffusion kurtosis imaging (DKI) diagnosis prognosis Figures Figure 1 Figure 2 Figure 3 Background H3K27 -altered diffuse midline gliomas (DMGs) are fatal malignancies of the central nervous system (CNS). According to the 2021 World Health Organization (WHO) classification, H3K27 -altered DMGs are classified as WHO Grade 4 tumors and characterized by diffuse and aggressive behavior[ 1 , 2 ].DMGs are categorized as a pediatric-type diffuse high-grade glioma in the CNS WHO 2021[ 2 ]. H3K27 -altered DMGs predominate in children but can also be seen in adolescents and adults[ 3 ]. These tumors constitute approximately 80% of brainstem tumors[ 4 ] and primarily develop within midline structures[ 5 ]. Previous studies usually compared adult DMGs with midline IDH wild-type glioblastomas (GBMs)[ 6 , 7 ]. However, in clinical practice, the differential diagnosis of brainstem tumors often involves a wider variety of glioma and glioneuronal subtypes, such as astrocytomas[ 8 ], pilocytic astrocytomas[ 9 ], and gangliogliomas[ 10 ], in addition to GBMs. They often present with overlapping clinical and imaging characteristics. Accurate differentiation of H3K27 -altered DMGs from H3K27 wide-type midline tumors (referred as non-DMGs) is crucial because of the significant variation in therapeutic strategies and prognostic implications. Magnetic resonance imaging (MRI) is the standard non-invasive examination for diagnosing DMGs and monitoring tumor response to treatment. Conventional MRI, combined with radiomics and deep learning, has been utilized to predict H3K27 genetic alterations[ 11 , 12 ] and prognosis[ 4 , 13 ]. However, these engineered features often lack interpretability, particularly in clinical practice. Furthermore, conventional MRI primarily provides anatomical insights into tumors and fails to capture the biological information related to H3K27 -altered DMGs. In the past decade, advanced MRI techniques such as amide proton transfer-weighted (APTw), arterial spin labeling (ASL), and multi-shell diffusion imaging have been utilized to characterize metabolism[ 14 , 15 ], angiogenesis[ 16 , 17 ], and cellular density[ 18 – 20 ] within brain tumors. Recently, several studies have applied these imaging techniques to explore the underlying pathology and aid in clinical diagnosis of DMGs. Zhuo et al.[ 15 ] demonstrated that APTw value and APTw image-derived radiomic features had good predictive ability for H3K27 mutation status, offering a novel imaging biomarker for the non-invasive assessment of DMGs. Studies also found that H3K27 -mutant tumors typically exhibit high cerebral blood flow (CBF), indicating increased angiogenic activity and invasiveness[ 21 , 22 ]. Furthermore, diffusion-weighted imaging (DWI) has been used to investigate the biological characteristics of DMGs, with lower apparent diffusion coefficient (ADC) values often associated with higher cellular density and poorer prognosis[ 23 ]. However, DWI is based on a Gaussian distribution model and cannot fully capture cellular complexity. In contrast, multi-shell diffusion MRI, such as diffusion kurtosis imaging (DKI), provides non-Gaussian diffusion information, offering a more comprehensive view of tissue complexity and microenvironmental changes[ 19 ]. A recent study using multi-shell diffusion MRI-derived metrics showed high discriminative power in distinguishing between H3K27 -altered and wild-type midline gliomas[ 24 ]. However, most previous studies have focused on identifying DMGs using a single imaging technique, failing to fully leverage the integration of metabolic, perfusion, and diffusion imaging for diagnosing H3K27 -altered DMGs. In addition, no study has used these MRI techniques to predict prognosis of H3K27 -altered DMGs, which is vital for treatment monitoring and clinical management. To better reflect this clinical reality, our study incorporated a broader spectrum of non-DMG midline tumors as the control group. This study aims to characterize the metabolic, perfusion, and diffusion characteristics of H3K27 -altered DMGs using APTw, ASL, and DKI. Furthermore, we systematically assess their added value for diagnosing H3K27 -altered DMGs and predicting prognosis, compared with conventional MRI. Methods Study Design The overview study design, including data collection, image processing, feature extraction, and statistical analysis, is presented in Fig. 1 . Ethics Approval This study was approved by the Animal and Human Ethics Committee of Beijing Tiantan Hospital, Capital Medical University (KY2022-078-02). Written informed consent was obtained from all patients or their legal guardians in accordance with the Declaration of Helsinki. Participants From June 2020 to May 2023, 125 patients with clinically and radiologically suspected brainstem tumors were prospectively recruited. Inclusion criteria were: (1) no prior treatment; (2) complete preoperative MRI examinations; (3) tumors were located in the brainstem, confirmed by MRI; and (4) scheduled for biopsy or surgical resection and with a final histopathological diagnosis. Exclusion criteria were: (1) poor image quality; and (2) incomplete histological information. Clinical information collected included age, sex, Karnofsky Performance Status (KPS) score at admission or clinic, and tumor histology. All patients underwent regular follow-up after diagnosis. Patients lost to follow-up were censored at their last known contact date. Overall survival (OS) was defined as the time from diagnosis to death from any cause. The final follow-up was on April 1, 2024. The median follow-up was 11.7 months (95% confidence interval [CI]: 3.7–28.8 months). Pathological Analysis The tumors were classified by an experienced pathologist (X.L, with 10 years of experience) according to the 2021 WHO classification of CNS tumors[ 2 ]. H3K27 alteration status was determined by immunohistochemical using a mutation-specific antibody (Merck Millipore, Billerica, MA, USA). The Ki-67 index was assessed by immunohistochemistry using the MIB-1 (anti-Ki-67) antibody (Dako, Denmark). Tumor cells were identified based on standard morphological criteria. In five hot spots (400× magnification), the total number of tumor cells and Ki-67- positive cells were manually counted. The Ki-67 index was calculated as the mean percentage of positive cells across these fields. MRI Acquisition MRI acquisition was performed on a 3T MR scanner (Ingenia CX, Philips Healthcare, Best, The Netherlands) with a 32-channel head coil. The MR protocol included T1-weighted image(T1w), T2-weighted image(T2w), and contrast-enhanced T1-weighted imaging(T1C). Details of the conventional sequences can be found in Additional file 1: Table S1 . For APTw, the protocol parameters for 3D fast spin-echo sequences were: repetition time (TR) = 5864 ms, echo time (TE) = 8.8 ms, flip angle (FA) = 90°, voxel size = 2 mm × 2 mm × 6 mm, with 7 slices acquired. Saturation pulses at + 3.5 ppm (relative to the water resonance frequency) were applied with a B1 amplitude of 2 µT for 2 s to saturate amide protons. To normalize and interpolate the Z-spectrum, six additional volumes were acquired at offset frequencies (± 3.1 ppm, − 3.5 ppm, ± 3.9 ppm, and − 1560 ppm). Three further acquisitions at + 3.5 ppm with varied echo-time shifts generated Dixon-type B0 field maps for correcting B0 inhomogeneity in the Z-spectrum frequency domain[ 25 ]. SENSE factor = 1.6, acquisition time = 1 min 54 s. Pseudo-continuous arterial spin labeling (pCASL) was used for ASL data acquisition. The acquisition parameters were as follows: 3D gradient and spin-echo imaging, 20 slices, 8 pairs of control/label, TR/TE = 3903 ms/11 ms, SENSE factor = 1.3, labeling duration = 1800 ms, post-labeling delay (PLD) = 1500 ms, scan time = 3 min 31 s. DKI was acquired using axial 2D spin-echo echo-planar imaging (SE-EPI). Acquisition parameters were as follows: TR/TE = 4000 ms/88 ms, FA = 90°, in-plane voxel size = 2.5 mm × 2.5 mm, acquisition matrix size = 96 × 96, slice thickness = 2.5 mm, number of slices = 60, b-values = 0, 1000, 2000 s/mm², with 48 diffusion-sensitizing gradient directions for each non-zero b-value, SENSE factor = 2, multiband factor = 3, acquisition time = 6 min 48 s. Additional phase-encoding reverse B0 images were acquired for distortion correction. Tumor Segmentation For tumor segmentation, the solid tumor was manually delineated on T2w images, excluding cystic, necrotic, and hemorrhagic areas. Segmentation was performed by two neuroradiologists (M.W. and J.Q., with over 5 and 10 years of neuroradiology experience, respectively), who were blinded to the H3K27 alteration status. T1w and T1C images served as anatomical references. Segmentation was executed using 3D Slicer software (version 4.10.2; www.slicer.org ). All segmentations were subsequently reviewed and, if necessary, refined by a senior neuroradiologist (Y.D., with 20 years of experience). The Dice similarity coefficient between the two raters for the solid tumor segmentation was 0.89. The intersection of the two reviewers’ segmentations was used as the final tumor mask. Tumor volume was calculated from this final segmentation. Image Processing APTw images were automatically processed using the embedded post-processing pipeline on the MR console. Magnetization transfer ratio asymmetry (MTRasym) analysis was conducted voxel-by-voxel to differentiate APT signals from background effects[ 26 ]. B0 field maps were used to correct for inhomogeneity artifacts, and the MTRasym value at + 3.5 ppm was calculated and expressed as a percentage. To extract APTw values, T2w images, and tumor segmentation were coregistered to the corresponding APTw images, and the mean APTw value within the solid tumor was measured for each case. Relative cerebral blood flow (rCBF) was calculated from ASL data using the embedded post-processing pipeline. To account for inter-individual variation in perfusion, the rCBF maps were normalized by subtracting the mean gray-matter rCBF and dividing by that same mean value. Cerebral gray matter was segmented from T1w images using Statistical Parametric Mapping (SPM12, https://www.fil.ion.ucl.ac.uk/spm/ ) and coregistered to the ASL source images. Next, T2w images and tumor segmentation mask were coregistered to the corresponding ASL source images. Finally, the mean normalized rCBF within the tumor mask was extracted for each case; \(\:rCBF=\frac{Lesional\:CBF-Cerebellar\:Hemispℎeric\:CBF}{Cerebellar\:Hemispℎeric\:CBF}\) ༛ Normalized rCBF is referred to simply as rCBF in subsequent analyses. DKI images were processed using DIPY ( https://dipy.org/ ), FMRIB Software Library (FSL) version 6.0 ( https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL ), and MRtrix3 ( https://www.mrtrix.org/ ). Preprocessing steps included denoising, distortion correction, eddy-current correction, and skull stripping. Seven diffusion metrics were calculated: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), mean kurtosis (MK), axial kurtosis (AK), and radial kurtosis (RK). To extract diffusion metrics, T2w images, and tumor segmentation mask were coregistered to the corresponding B0 images. Finally, the mean diffusion metrics within the solid tumor were measured for each patient. Radiological Evaluation Two independent neuroradiologists (M.W. and J.Q., with over 5 and 10 years of neuroradiology experience, respectively) characteristics on conventional MRI sequences (T1w, T2w, and T1C), including: (1) tumor location (midbrain, pons, or medulla); (2) presence of hydrocephalus; (3) contrast enhancement patterns; and (4) necrotic. Hydrocephalus was diagnosed based on MRI findings of the lateral ventricles (Evans index > 0.3)[ 27 ]. The extent of resection (EOR) on postoperative images was classified as gross total resection (GTR: 100% EOR), subtotal resection (STR: > 50% and < 100% EOR), and partial resection (PR: < 50% EOR). Discrepancies were resolved by a senior neuroradiologist (Y.D. with over 20 years of neuroradiology experience), who reviewed and confirmed the final assessments. Statistical Analysis Data analysis was performed using R statistical software (version 4.2.3). Categorical variables were reported as frequencies and percentages and analyzed by the Chi-square test. Continuous variables were compared using Student's t-test or the Mann-Whitney U test, depending on data distribution. First, we assessed the separate and combined contributions of APTw, ASL, and DKI-derived measures to the diagnosis of DMGs. Multivariate logistic regression with leave-on-out cross-validation (LOOCV) was used to distinguish H3K27 -altered DMGs from non-DMGs using five models (Model 1: clinical variables and conventional MR features; Model 2: Model 1 + APTw; Model 3: Model 1 + ASL; Model 4: Model 1 + DKI; and Model 5: Model 1 + APTw + ASL + DKI). We then calculated the area under the curve (AUC), sensitivity, specificity, and accuracy to evaluate model performance. Differences between models were assessed using DeLong’s test. Second, we assessed the separate and integrated contributions of APTw, ASL, and DKI metrics to the prognosis of DMGs. Multivariate Cox regression with LOOCV was used to build five prognostic models (similar as above five models in diagnostic tasks) to predict OS in H3K27 -altered DMGs group. Model performance was assessed by the concordance index (C-index). Additional survival analyses stratified by these metrics and risk predictions by Cox regression (Cox model 5) were conducted using the Kaplan-Meier method and log-rank test. A p-value < 0.05 was considered statistically significant. Results Demographic, Clinical, and Conventional MRI characteristics A total of 125 patients with brainstem glioma were initially eligible. Four patients were excluded for prior treatment before biopsy or surgery. Six for poor APTw image quality. Five for poor ASL image quality. Ten patients for poor DKI image quality and five for missing histological information (Additional file 1: Fig. S1 ). Ultimately, 95 patients were included: 71 patients with H3K27 -altered DMGs and 24 patients with non-DMGs. The non-DMG group comprised seven pilocytic astrocytomas, two gangliogliomas, seven wild-type GBMs, six astrocytomas (including two Grade 4, one Grade 3, and three Grade 2), one mixed glioma-neuronal tumor, and one gliosis. Representative conventional MRI and quantitative parameter maps (APTw for metabolism, ASL-CBF for perfusion, DKI-derived MD for diffusion) from two representative cases are illustrated in Fig. 2 A. Patients with H3K27 -altered DMGs were younger than those with non-DMGs (15.13 ± 1.58 years vs. 22.46 ± 3.60 years, p = 0.035); the age distribution of patients is shown in Additional file 1: Fig. S2 . Mean KPS was lower in H3K27 -altered DMG group than in non-DMGs group (86.06 vs. 97.08, p = 0.012). Most H3K27 -altered DMGs were located in the pons (66/71; 92.96%). Non-DMGs were located in the pons (62.50%), medulla (8.33%), and midbrain (29.17%). Significant differences in EOR were observed between the H3K27-altered DMG and non-DMG groups: biopsy (40.85% vs. 16.67%); GTR (16.90% vs. 62.50%); STR (9.86% vs. 16.67%); PR (32.39% vs. 4.16%) (all p < 0.001) (Table 1 ). There were no differences between H3K27 -altered DMG and non-DMG groups in conventional MRI features (e.g., enhancement, necrosis, hydrocephalus) (Table 1 ). Table 1 Demographics, clinical, and conventional MRI characteristics of patients with DMGs and non-DMGs All Patients ( N = 95 ) H3K27 -altered DMGs ( N = 71 ) non-DMGs ( N = 24 ) p-value Age 16.98 ± 1.52 15.13 ± 1.58 22.46 ± 3.60 0.035* Sex (%) 1.000 male 37 (38.95% ) 28 (39.44% ) 9 (37.50% ) female 58 (61.05% ) 43 (60.56% ) 15 (62.50% ) EOR (%) <0.001* Biopsy 33 (34.74% ) 29 (40.85% ) 4 (16.67% ) GTR 27 (28.42% ) 12 (16.90% ) 15 (62.50% ) STR 11 (11.58% ) 7 (9.86% ) 4 (16.67% ) PR 24 (25.26% ) 23 (32.39% ) 1 (4.16% ) KPS 88.84 ± 1.81 86.06 ± 2.30 97.08 ± 1.27 0.012* Location (%) <0.001* Pons 81 (85.26% ) 66 (92.96% ) 15 (62.50% ) Medulla 7 (7.37% ) 5 (7.04% ) 2 (8.33% ) Midbrain 7 (7.37% ) 0 (0.00% ) 7 (29.17% ) Enhancement (%) 1.000 no 35 (36.84% ) 26 (36.62% ) 9 (37.50% ) yes 60 (63.16% ) 45 (63.38% ) 15 (62.50% ) Necrosis (%) 0.767 no 48 (50.53% ) 37 (52.11% ) 11 (45.83% ) yes 47 (49.47% ) 34 (47.89% ) 13 (54.17% ) Hydrocephalus (%) 1.000 no 65 (68.42% ) 49 (69.01% ) 16 (66.67% ) yes 30 (31.58% ) 22 (30.99% ) 8 (33.33% ) Volume (cm 3 ) 22.7 [15.3;29.7] 23.4 [19.0;30.0] 16.4 [10.4;24.5] 0.008* APTw 2.74 [2.07;3.45] 2.79 [2.30;3.50] 2.01 [1.63;2.74] 0.001* rCBF -0.35 [-0.50;-0.18] -0.30 [-0.46;-0.17] -0.49 [-0.61;-0.34] 0.002* FA 0.23 [0.20;0.27] 0.24 [0.20;0.28] 0.20 [0.18;0.25] 0.015* MD (×10 − 3 mm/s) 1.34 [1.09;1.60] 1.24 [1.02;1.49] 1.43 [1.34;1.78] 0.002* AD (×10 − 3 mm/s) 1.65 [1.38;1.91] 1.57 [1.32;1.83] 1.78 [1.65;2.05] 0.002* RD (×10 − 3 mm/s) 1.19 [0.96;1.44] 1.08 [0.86;1.35] 1.30 [1.18;1.63] 0.001* MK 0.58 [0.51;0.68] 0.58 [0.52;0.67] 0.57 [0.46;0.72] 0.837 AK 0.57 [0.49;0.63] 0.57 [0.50;0.62] 0.54 [0.46;0.66] 0.461 RK 0.61 [0.54;0.75] 0.61 [0.55;0.73] 0.64 [0.47;0.80] 0.932 EOR = extent of resection; GTR = gross-total resection; STR = subtotal resection, PR = partial resection; KPS = Karnofsky Performance Scale; APTw = amide proton transfer-weighted; rCBF = relative cerebral blood flow; FA = fractional anisotropy; MD = mean diffusivity; RD = radial diffusivity; AD = axial diffusivity; AK = axial kurtosis; MK = mean kurtosis; RK = radial kurtosis; HR = Hazard ratio. * indicate p < 0.05 Metabolic, Perfusion, and Diffusion Characteristics of DMGs For metabolic, perfusion, and diffusion metrics, patients with H3K27 -altered DMGs had significantly different values compared with non-DMG patients for APTw (median 2.79 vs. 2.01); rCBF (median − 0.30 vs. -0.49); FA (median 0.24 vs. 0.20); MD (median 1.24 vs. 1.43); AD (median 1.57 vs. 1.78); and RD (median 1.08 vs. 1.30) (all p < 0.05). However, DKI metrics (MK, AK, and RK) did not show significant differences (Table 1 , Fig. 2 B). Metabolic, Perfusion, and Diffusion Imaging Improve Diagnostic Performance for H3K27 -Altered DMGs We Constructed five multivariate logistic regression models using variables that were significant in univariate analyses. Univariate logistic regression analysis results are shown in Additional file 1: Table S2 . Due to high collinearity among diffusion metrics. Because MD equals the average of AD and RD, it serves as a composite metric reflecting both diffusion components[ 28 ]. So, MD was selected for multivariable logistic and Cox regression analyses (Table 2 ). Table 2 Five multivariable regression models to identify H3K27-altered DMGs from non-DMGs Model 1 Model 2 Model 3 Model 4 Model 5 OR (95% CI) p OR (95% CI) p OR (95% CI) p OR (95% CI) p OR (95% CI) p Location Pons ref ref ref ref ref Medulla 2.23(0.32,22.25) 0.447 2.19(0.24,26.99) 0.505 2.53(0.31,29.85) 0.414 1.53(0.17,20.00) 0.722 1.41(0.11,23.26) 0.798 Midbrain NA NA NA NA NA Age 0.97(0.93,1.01) 0.183 0.98(0.94,1.03) 0.341 0.96(0.96,1.00) 0.079 0.95(0.90,0.99) 0.029* 0.95(0.89,1.00) 0.043* KPS 0.91(0.81,0.98) 0.058 0.92(0.81,0.98) 0.064 0.92(0.81,0.99) 0.099 0.90(0.77,0.98) 0.066 0.94(0.82,1.01) 0.203 Volume 1.03(0.97,1.09) 0.411 1.02(0.96,1.08) 0.614 1.00(0.94,1.07) 0.991 1.05(0.99,1.14) 0.154 1.02(0.95,1.10) 0.578 APTw 3.81(1.65,10.65) 0.004* 3.41(1.39,10.28) 0.014* rCBF 7.74(3.18,44.25) 0.019* 1.26(0.25,18.88) 0.265 MD 0.03(0.001,0.38) 0.012* 0.05(0.002,0.80) 0.049* FA 0.88(0.16,4.57) 0.877 0.98(0.14,6.36) 0.982 KPS = Karnofsky Performance Scale; APTw = amide proton transfer-weighted; rCBF = relative cerebral blood flow ;FA = fractional anisotropy; MD = mean diffusivity; OR = odds ratios; CI = confidence interval; ref = reference; NA, not available. * indicates p < 0.05. Model 1 included clinical variables and conventional MRI features. Model 2 added APTw measures to Model 1 and achieved an odds ratio (OR) of 3.81 (95% CI: 1.65–10.65, p = 0.004) with an AUC was 0.901, a 0.07 improvement over Model 1 (Delong’s test p = 0.125). Model 3 added rCBF measures to Model 1 and achieved an OR of 7.74 (95% CI: 3.18–44.25, p = 0.019) with an AUC was 0.881, a 0.05 improvement over Model 1 (Delong’s test p = 0.216). Model 4 added DKI metrics (MD and FA) to Model 1, MD remained an independent predictor (an OR of 0.03; 95% CI: 0.001–0.38, p = 0.012), and an AUC of Model was 0.904, a 0.073 improvement over Model 1 (Delong’s test p = 0.046). Model 5 integrated APTw, rCBF, MD, and FA into Model 1, APTw and MD remained significant independent contributors, with ORs of 3.41 (95% CI: 1.39–10.28, p = 0.014) and 0.05 (95% CI: 0.002- 0.80, p = 0.049), respectively. The AUC for Model 5 was 0.994, a 0.163 improvement over Model 1 (Delong’s test p = 0.030). Model 5 (clinical variables + conventional MRI + APTw + ASL + DKI) achieved the highest AUC among all models, indicating the best diagnostic performance (Fig. 3 A). The AUC and cutoff values for the diagnostic prediction of each variable are presented in Additional file 1: Table S3. Performance of five models for the diagnostic prediction in Additional file 1: Table S4. Delong’s test result is in Additional file 1: Fig. S3. The AUC values of the ROC curves obtained by using each imaging technique alone (APTw, ASL, DKI) have been provided in Additional file 1: Fig. S4 and Additional file 1: Table S5. Metabolic, Perfusion, and Diffusion Imaging Improve Prognostic Prediction for H3K27 -Altered DMGs Among the 71 DMG patients with available OS data, we constructed five multivariate Cox regression models mirroring the diagnostic analysis framework (Table 3 ). Variables achieving statistical significance (p < 0.05) on univariate COX regression analyses were selected for multivariable analysis. Table 3 Multivariable Cox regression models for predicting OS in H3K27 -altered DMGs Model 1 Model 2 Model 3 Model 4 Model 5 HR (95% CI) p HR (95% CI) P HR (95% CI) p HR (95% CI) p HR (95% CI) p Location Pons ref ref ref ref ref Medulla 3.99(0.90, 17.60) 0.068 3.56(0.79, 16.0) 0.100 4.87(1.09, 21.81) 0.038 6.25(1.10, 35.70) 0.039 4.78(0.77, 29.6) 0.093 Age 1.00(0.98, 1.02) 0.876 1.00(0.97, 1.02) 0.737 1.00(0.97, 1.02) 0.805 1.00(0.97, 1.02) 0.788 0.99(0.97, 1.02) 0.630 KPS 1.00(0.98, 1.02) 0.802 1.00(0.99, 1.02) 0.619 1.01(0.99, 1.03) 0.583 1.00(0.98, 1.02) 0.932 1.01(0.99, 1.03) 0.521 EOR Biopsy ref ref ref ref ref GTR 0.28(0.08, 0.93) 0.038* 0.21(0.06, 0.71) 0.012* 0.27(0.08, 0.90) 0.033* 0.23(0.06, 0.81) 0.023* 0.20(0.05, 0.72) 0.014* STR 0.52(0.15, 1.72) 0.286 0.45(0.13, 1.55) 0.230 0.51(0.15, 1.75) 0.332 0.64(0.18, 2.31) 0.482 0.49(0.13, 1.86) 0.294 PR 0.92(0.43, 1.97) 0.781 0.72(0.33, 1.59) 0.386 0.79(0.35, 1.77) 0.615 0.96(0.45, 2.06) 0.876 0.72(0.32, 1.62) 0.428 Ki-67 1.04(1.02, 1.06) < 0.001* 1.03(1.01, 1.05) 0.015* 1.05(1.03, 1.07) < 0.001* 1.04(1.02, 1.06) < 0.001* 1.04(1.01, 1.06) 0.003* Volume 1.03(1.00, 1.05) 0.082 1.02(0.99, 1.05) 0.185 1.02(0.99, 1.05) 0.173 1.02(0.99, 1.06) 0.110 1.02(0.99, 1.05) 0.257 APTw 1.89(1.24, 2.87) 0.003* 1.66(1.03, 2.67) 0.037* rCBF 2.87(1.46, 5.65) 0.002* 2.15(1.05, 4.44) 0.038* MD 1.69(0.45, 6.34) 0.382 1.15(0.25, 5.25) 0.856 FA 1.72(0.62, 4.80) 0.336 1.21(0.40, 3.69) 0.733 OS = overall survival; KPS = Karnofsky Performance Scale; EOR = extent of resection; GTR = gross-total resection; STR = subtotal resection, PR = partial resection; APTw = amide proton transfer-weighted; rCBF = relative cerebral blood flow; FA = fractional anisotropy; MD = mean diffusivity; HR = Hazard ratios; CI = confidence interval; ref = reference. * indicate p < 0.05. Model 1 included clinical variables and conventional MRI features. Model 2 added APTw to Model 1, APTw remained an independent predictor (hazard ratio [HR] of 1.89; 95% CI: 1.24–2.87; p = 0.003). The C-index of Model 2 was 0.714, a 0.009 decrease from Model 1 (C-index = 0.723). Model 3 added rCBF to Model 1, it yielded an HR of 2.87 (95% CI: 1.46–5.65, p = 0.002). The C-index of Model 3 was 0.745, a 0.022 improvement over Model 1. Model 4 incorporating DKI metrics (MD and FA), an HR of MD was 1.69 (95% CI: 0.45–6.34, p = 0.382). The HR of FA was 1.72 (95% CI: 0.62–4.80, p = 0.336). The C-index of Model 4 was 0.729, a 0.006 increase over Model 1, but neither MD nor FA reached statistical significance. Model 5 integrated APTw, rCBF, MD, and FA, APTw remained an independent predictor with an HR of 1.66 (95% CI: 1.03–2.67, p = 0.037), and rCBF also maintained its significance with an HR of 2.15 (95% CI: 1.05–4.44, p = 0.038). The C-index of Model 5 was 0.742, an improvement over Model 1. Additionally, Ki-67 index and GTR were significant prognostic factors in all models (Table 3 ). Evaluation metrics for the five models are provided in Additional file 1: Table S6; univariate Cox results in Additional file 1: Table S7; Fig. 3 B shows the forest plot of Model 5. We further stratified patients by median APTw and rCBF values, log-rank p-values were 0.002 and 0.010, with HR values of 2.55 and 2.13, respectively (Fig. 3 C, 3 D). Predicted risks from Model 5 classified patients into high-risk and low-risk groups. The Kaplan-Meier curve for Model 5 showed an HR value of 8.57 (p = 0.003) between the risk groups (Fig. 3 E). To explore histological association of APTw, rCBF, and MD significantly that enhanced the diagnostic and prognostic performances of H3K27 -altered DMGs, we correlated APTw, rCBF, and MD with the Ki-67 index. The results showed a significant correlation between Ki-67 and APTw (Additional file 1: Fig. S5), whereas rCBF and MD did not. Discussion In this study, we used metabolic, perfusion, and diffusion imaging to assess the diagnostic performance and prognostic value of H3K27 -altered DMGs. The main findings were as follows: (1) Patients with H3K27 -altered DMGs were significantly younger and had lower KPS scores, and they showed significant differences in APTw, rCBF, FA, MD, AD, and RD values compared with non-DMGs. (2) APTw, rCBF, and MD each had independent predictive value for diagnosing H3K27 -altered DMGs, and the combination of APTw, rCBF, and MD (Model 5) achieved the highest diagnostic performance, with an AUC of 0.944, a 0.163 improvement compared with conventional MRI. (3) Both APTw and rCBF were independent prognostic factors for OS in H3K27 -altered DMGs, and the combination of APTw, rCBF, MD, and clinical variables provided the most robust prognostic performance. H3K27 -altered DMGs are pediatric-type diffuse high-grade gliomas defined in the 2021 WHO classification[ 2 ]. We found that patients with H3K27 -altered DMGs were significantly younger than those without DMGs, reflecting tumor development in the developing brain[ 29 ]. These patients also had lower KPS scores. Increased APTw values likely reflect active tumor cell proliferation[ 30 ]. Compared with non-DMGs, H3K27 -altered DMGs exhibited higher rCBF values, which may relate to a positive correlation between rCBV and H3K27M -positive nuclear density in tumor samples[ 22 ]. H3K27 -altered DMGs also showed increased FA values and reduced diffusion metrics (MD, AD, and RD) values. FA measures diffusion anisotropy, indicating greater tissue complexity, while diffusion metrics (MD, AD, and RD) assess directional diffusion, reflecting restricted water movement in tissue[ 20 ]. These findings suggest higher tumor cell density and more complex tissue microstructure in H3K27 -altered DMGs. In diagnostic models of H3K27 -altered DMGs, APTw imaging assesses heterogeneous protein and peptide metabolism, which may reflect the histopathology and genetic alterations of gliomas[ 14 , 15 , 30 ]. Higher APTw values indicate increased protein and peptide metabolism and likely reflect active tumor cell proliferation[ 30 ]. In our study, adding APTw values to clinical and conventional imaging features in the multivariable logistic regression model significantly improved diagnostic performance for H3K27 -altered DMGs. This finding aligns with Zhuo et al.[ 15 ], who showed that APTw-derived radiomics features predicted H3K27 mutations in brainstem glioma patients with 86% accuracy. Tumor perfusion also contributed to diagnostic accuracy for H3K27 -altered DMGs. Our study showed that rCBF had significant diagnostic power in the multivariable model and independently predicted H3K27 -altered DMGs. This finding agrees with Kathrani et al.[ 21 ], who found that rCBV ratios independently correlate with histone mutation status in H3K27 -altered DMGs. We found that diffusion metrics (MD, AD, and RD) significantly predicted H3K27 -altered DMGs and enhanced diagnostic performance, potentially reflecting increased tumor cell density 16 . Xu et al.[ 20 ] made similar findings, suggesting that the densely packed tumor cells water diffusion, resulting in lower diffusion metrics values. This reduction may relate to tumor heterogeneity, including cellular proliferation and microenvironment changes[ 19 , 20 ]. The combination of APTw, rCBF, and MD achieved the highest diagnostic performance for H3K27 -altered DMGs. Integrating these metrics with clinical and conventional imaging features yielded an AUC of 0.944, a 0.163 improvement compared with conventional MRI, demonstrating the added diagnostic value of combining these metabolic, perfusion, and diffusion parameters. In the models predicting OS of H3K27 -altered DMGs, APTw independently predicted OS, and patients with high APTw values had significantly lower OS (p = 0.002). Supporting these findings, Joo et al.[ 30 ] also found that higher APT signals correlate with poorer prognosis in high-grade glioma patients. Previous studies suggest that higher APT signals reflect increased tumor cell proliferation and invasiveness and are associated with poor prognosis[ 31 , 32 ]. Our results confirm a certain significant correlation between the APTw values and the Ki-67 index, with a high Ki-67 index indicating rapid tumor cell proliferation[ 33 ]. In H3K27 -altered DMGs, the Ki-67 index (Ki-67 ≤ 5% vs. >5%) showed prognostic value[ 34 ], and several studies report differences in Ki-67 proliferation among DMG patients that correlate with biological behavior and prognosis[ 35 , 36 ]. Therefore, we investigated correlations between Ki-67 and imaging parameters to assess its potential as an auxiliary prognostic indicator. The results showed that rCBF also had significant prognostic ability and was an independent predictor of OS in H3K27 -altered DMGs, with high rCBF values associated with shorter OS. Pang et al.[ 37 ] also found that high CBF values are associated with poor prognosis in glioma patients. They observed that CBF correlates with VEGF expression, suggesting that CBF may approximate angiogenesis across glioma grades. We observed that although diffusion metrics (MD, AD, and RD) were valuable for the diagnosis and classification of H3K27 -altered DMGs, they did not significantly contribute to OS prediction, indicating that they may not fully cover factors influencing prognosis. Consequently, combining APTw, rCBF, and MD, along with clinical variables yielded the most robust prognostic performance by leveraging their complementary strengths. Furthermore, APTw and rCBF remained the primary independent predictors of OS, underscoring their value in prognostication for H3K27 -altered DMG patients. Our results demonstrate the value of metabolic, perfusion, and diffusion imaging for diagnosing and predicting prognosis in H3K27 -altered DMGs. Nevertheless, this study has certain limitations. First, our cohort included 95 patients (71 H3K27 -altered DMGs and 24 non-DMGs), resulting in class imbalance and a relatively small sample size. Although we used LOOCV for internal validation, this does not fully eliminate bias from sample size and class imbalance. Future studies should verify these results in larger, multicenter cohorts to improve the generalizability. Second, incomplete records of treatments beyond EOR prevented adjustment for all therapeutic confounders, potentially biasing the results; larger, multicenter cohorts with detailed treatment data are needed to address this issue. Third, the median follow-up of 11.7 months may be insufficient to assess long-term survival; future studies should include longer follow-up to verify the stability of these findings. Finally, we did not perform histopathological or longitudinal imaging validation to link imaging metrics with tumor biology over time. Future work should integrate histology, repeated imaging, or biopsy data within long-term cohorts to assess the biological significance and temporal stability of these biomarkers. Conclusion This study improved diagnostic accuracy and prognostic prediction of H3K27 -altered DMGs by integrating metabolic, perfusion, and diffusion metrics with conventional MRI and clinical variables. The results demonstrate that integrating metabolic, perfusion, and diffusion imaging with conventional MRI and clinical variables provides robust support for accurate diagnosis and prognostic assessment of H3K27 -altered DMGs. Abbreviations AD axial diffusivity ADC apparent diffusion coefficient AK axial kurtosis APTw amide proton transfer-weighted ASL arterial spin labeling AUC area under the curve DMGs diffuse midline gliomas non-DMGs non- H3K27 -altered DMGs CBF cerebral blood flow CI confidence interval CNS central nervous system DKI diffusion kurtosis imaging DWI diffusion-weighted imaging EOR extent of resection GTR gross-total resection FA fractional anisotropy HR Hazard ratio KPS Karnofsky Performance Scale LOOCV leave-on-out-cross-validation MD mean diffusivity MK mean kurtosis MRI Magnetic resonance imaging MTRasym Magnetization transfer ratio asymmetry OR odds ratios OS overall survival pCASL pseudo-continuous Arterial Spin Labeling PLD post-labeling delay PR partial resection rCBF relative cerebral blood flow RD radial diffusivity RK radial kurtosis ROC receiver operating characteristic SE-EPI spin-echo echo-planar imaging STR subtotal resection T1w T1-weighted T1C contrast-enhanced T1-weighted T2w T2-weighted TE echo time TR repetition time WHO World Health Organization (WHO) Declarations Ethics approval and consent to participate This study was approved by the Animal and Human Ethics Committee of Beijing Tiantan Hospital, Capital Medical University (KY2022-078-02). Written informed consent was obtained from all patients or their legal guardians in accordance with the Declaration of Helsinki. Consent for publication All authors approved the submitted version of the paper. Availability of data and materials The corresponding authors had full control of the study data, and all data generated or analyzed during the study are available from the corresponding author by request. Competing interests The authors declare that they have no competing interests. Funding This work was supported by Beijing Postdoctoral Foundation (NO.2-1-2-006-46); Beijing Natural Science Foundation (Grant Nos. 7252040, 7244328); Capital Health Development Scientific Research Special Project of Beijing Municipal Health Commission (NO.2022-1-2042); the Nursery engineering project of Beijing Tiantan Hospital Affiliated to Capital Medical University (No. 2023MP02); Capital Medical University (CCMU2024ZKYXZ007). Author contributions: Jun Qiu: Conceptualization, Formal analysis, Investigation and Writing - Original Draft. Junjie Li: Visualization, Investigation. Yunyun Duan : Investigation and Writing - Review & Editing. Jun Sun : Software and Data Curation; Yuna Li : Resources and Formal analysis. Min Guo : Resources. Minghao Wu : Resources; Xiaolu Xu : Investigation. Tiantian Hua : Resources; Z hizheng Zhuo: Project administration; Yuwei Liu: Supervision; Ying Jin: interpreted the data; Xing Liu: technical support; Liwei Zhang: Study design; Zhizheng Zhuo: Supervision and Project administration, Writing - Review & Editing. Yaou Liu: Conceptualization, Supervision, Project administration, Funding acquisition, and Review & Editing. Authors' information: * Jun Qiu [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Department of Radiology, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui, 230036, China Junjie Li [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Yunyun Duan [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Jun Sun [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Yuna Li [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Min Guo [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Minghao Wu [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Xiaolu Xu [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Tiantian Hua [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Yuwei Liu [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Ying Jin [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China Xing Liu [email protected] Department of Pathology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China # Liwei Zhang Liwei Zhang [email protected] Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing 10070, China # Zhizheng Zhuo [email protected] Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China # Yaou Liu (Corresponding author) [email protected] 010-83911069 Department of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China References Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison DW. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol. 2016;131(6):803–20. Louis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D, Hawkins C, Ng HK, Pfister SM, Reifenberger G, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol. 2021;23(8):1231–51. Hayashi N, Fukai J, Nakatogawa H, Kawaji H, Yoshioka E, Kodama Y, Nakajo K, Uda T, Naito K, Kijima N, et al. Neuroradiological, genetic and clinical characteristics of histone H3 K27-mutant diffuse midline gliomas in the Kansai Molecular Diagnosis Network for CNS Tumors (Kansai Network): multicenter retrospective cohort. Acta Neuropathol Commun. 2024;12(1):120. Leach JL, Roebker J, Schafer A, Baugh J, Chaney B, Fuller C, Fouladi M, Lane A, Doughman R, Drissi R, et al. MR imaging features of diffuse intrinsic pontine glioma and relationship to overall survival: report from the International DIPG Registry. Neuro Oncol. 2020;22(11):1647–57. Lovibond S, Gewirtz AN, Pasquini L, Krebs S, Graham MS. The promise of metabolic imaging in diffuse midline glioma. Neoplasia. 2023;39:100896. Raab P, Banan R, Akbarian A, Esmaeilzadeh M, Samii M, Samii A, Bertalanffy H, Lehmann U, Krauss JK, Lanfermann H et al. Differences in the MRI Signature and ADC Values of Diffuse Midline Gliomas with H3 K27M Mutation Compared to Midline Glioblastomas. Cancers (Basel) 2022, 14(6). Sim Y, Choi SH, Lee N, Park YW, Ahn SS, Chang JH, Kim SH, Lee SK. Clinical, qualitative imaging biomarkers, and tumor oxygenation imaging biomarkers for differentiation of midline-located IDH wild-type glioblastomas and H3 K27-altered diffuse midline gliomas in adults. Eur J Radiol. 2024;173:111384. Chen LH, Pan C, Diplas BH, Xu C, Hansen LJ, Wu Y, Chen X, Geng Y, Sun T, Sun Y, et al. The integrated genomic and epigenomic landscape of brainstem glioma. Nat Commun. 2020;11(1):3077. Pan C, Li T, Zhang M, Wang Y, Xiao X, Zhang P, Zuo P, Wang Y, Xiao D, Wu Z, et al. A novel radiological classification of midbrain pilocytic astrocytomas and its implication for surgical management: a single-institution experience of 57 cases. J Neurosurg. 2024;140(6):1527–39. Acharya S, Liu JF, Tatevossian RG, Chiang J, Qaddoumi I, Gajjar A, Walker D, Harreld JH, Merchant TE, Ellison DW. Risk stratification in pediatric low-grade glioma and glioneuronal tumor treated with radiation therapy: an integrated clinicopathologic and molecular analysis. Neuro Oncol. 2020;22(8):1203–13. Chauhan RS, Kulanthaivelu K, Kathrani N, Kotwal A, Bhat MD, Saini J, Prasad C, Chakrabarti D, Santosh V, Uppar AM, et al. Prediction of H3K27M mutation status of diffuse midline gliomas using MRI features. J Neuroimaging. 2021;31(6):1201–10. Pan CC, Liu J, Tang J, Chen X, Chen F, Wu YL, Geng YB, Xu C, Zhang X, Wu Z, et al. A machine learning-based prediction model of H3K27M mutations in brainstem gliomas using conventional MRI and clinical features. Radiother Oncol. 2019;130:172–9. Liu X, Jiang Z, Roth HR, Anwar SM, Bonner ER, Mahtabfar A, Packer RJ, Kazerooni AF, Bornhorst M, Linguraru MG. Early prognostication of overall survival for pediatric diffuse midline gliomas using MRI radiomics and machine learning: A two-center study. Neurooncol Adv. 2024;6(1):vdae108. von Knebel Doeberitz N, Kroh F, Breitling J, Konig L, Maksimovic S, Grass S, Adeberg S, Scherer M, Unterberg A, Bendszus M, et al. CEST imaging of the APT and ssMT predict the overall survival of patients with glioma at the first follow-up after completion of radiotherapy at 3T. Radiother Oncol. 2023;184:109694. Zhuo Z, Qu L, Zhang P, Duan Y, Cheng D, Xu X, Sun T, Ding J, Xie C, Liu X, et al. Prediction of H3K27M-mutant brainstem glioma by amide proton transfer-weighted imaging and its derived radiomics. Eur J Nucl Med Mol Imaging. 2021;48(13):4426–36. Hou H, Chen W, Diao Y, Wang Y, Zhang L, Wang L, Xu M, Yu J, Song T, Liu Y, et al. 3D Amide Proton Transfer-Weighted Imaging for Grading Glioma and Correlating IDH Mutation Status: Added Value to 3D Pseudocontinuous Arterial Spin Labelling Perfusion. Mol Imaging Biol. 2023;25(2):343–52. Qiu J, Wu K, Zhu M, Chen CY, Luo Y, Liu Y, Wen J. Using quantitative MRI to study the association of isocitrate dehydrogenase (IDH) status with oxygen metabolism and cellular structure changes in glioma. Eur J Radiol. 2022;155:110502. Figini M, Riva M, Graham M, Castelli GM, Fernandes B, Grimaldi M, Baselli G, Pessina F, Bello L, Zhang H, et al. Prediction of Isocitrate Dehydrogenase Genotype in Brain Gliomas with MRI: Single-Shell versus Multishell Diffusion Models. Radiology. 2018;289(3):788–96. Zhao J, Wang YL, Li XB, Hu MS, Li ZH, Song YK, Wang JY, Tian YS, Liu DW, Yan X, et al. Comparative analysis of the diffusion kurtosis imaging and diffusion tensor imaging in grading gliomas, predicting tumour cell proliferation and IDH-1 gene mutation status. J Neurooncol. 2019;141(1):195–203. Xu X, Zhang P, Zhuo Z, Duan Y, Qu L, Cheng D, Sun T, Ding J, Xie C, Liu X et al. Prediction of H3K27M Alteration Status in Brainstem Glioma Using Multi-Shell Diffusion MRI Metrics. J Magn Reson Imaging 2023. Kathrani N, Chauhan RS, Kotwal A, Kulanthaivelu K, Bhat MD, Saini J, Prasad C, Chakrabarti D, Santosh V, Uppar AM, et al. Diffusion and perfusion imaging biomarkers of H3 K27M mutation status in diffuse midline gliomas. Neuroradiology. 2022;64(8):1519–28. Calmon R, Dangouloff-Ros V, Varlet P, Deroulers C, Philippe C, Debily MA, Castel D, Beccaria K, Blauwblomme T, Grevent D, et al. Radiogenomics of diffuse intrinsic pontine gliomas (DIPGs): correlation of histological and biological characteristics with multimodal MRI features. Eur Radiol. 2021;31(12):8913–24. Leibetseder A, Leitner J, Mair MJ, Meckel S, Hainfellner JA, Aichholzer M, Widhalm G, Dieckmann K, Weis S, Furtner J, et al. Prognostic factors in adult brainstem glioma: a tertiary care center analysis and review of the literature. J Neurol. 2022;269(3):1574–90. Zhang H, Liu K, Ba R, Zhang Z, Zhang Y, Chen Y, Gu W, Shen Z, Shu Q, Fu J, et al. Histological and molecular classifications of pediatric glioma with time-dependent diffusion MRI-based microstructural mapping. Neuro Oncol. 2023;25(6):1146–56. Togao O, Keupp J, Hiwatashi A, Yamashita K, Kikuchi K, Yoneyama M, Honda H. Amide proton transfer imaging of brain tumors using a self-corrected 3D fast spin-echo dixon method: Comparison With separate B(0) correction. Magn Reson Med. 2017;77(6):2272–9. Zhou J, Heo HY, Knutsson L, van Zijl PCM, Jiang S. APT-weighted MRI: Techniques, current neuro applications, and challenging issues. J Magn Reson Imaging. 2019;50(2):347–64. Toma AK, Holl E, Kitchen ND, Watkins LD. Evans' index revisited: the need for an alternative in normal pressure hydrocephalus. Neurosurgery. 2011;68(4):939–44. Arab A, Wojna-Pelczar A, Khairnar A, Szabo N, Ruda-Kucerova J. Principles of diffusion kurtosis imaging and its role in early diagnosis of neurodegenerative disorders. Brain Res Bull. 2018;139:91–8. Liu I, Jiang L, Samuelsson ER, Marco Salas S, Beck A, Hack OA, Jeong D, Shaw ML, Englinger B, LaBelle J, et al. The landscape of tumor cell states and spatial organization in H3-K27M mutant diffuse midline glioma across age and location. Nat Genet. 2022;54(12):1881–94. Joo B, Han K, Ahn SS, Choi YS, Chang JH, Kang SG, Kim SH, Zhou J, Lee SK. Amide proton transfer imaging might predict survival and IDH mutation status in high-grade glioma. Eur Radiol. 2019;29(12):6643–52. Park JE, Kim HS, Park KJ, Kim SJ, Kim JH, Smith SA. Pre- and Posttreatment Glioma: Comparison of Amide Proton Transfer Imaging with MR Spectroscopy for Biomarkers of Tumor Proliferation. Radiology. 2016;278(2):514–23. Regnery S, Adeberg S, Dreher C, Oberhollenzer J, Meissner JE, Goerke S, Windschuh J, Deike-Hofmann K, Bickelhaupt S, Zaiss M, et al. Chemical exchange saturation transfer MRI serves as predictor of early progression in glioblastoma patients. Oncotarget. 2018;9(47):28772–83. Uxa S, Castillo-Binder P, Kohler R, Stangner K, Muller GA, Engeland K. Ki-67 gene expression. Cell Death Differ. 2021;28(12):3357–70. Zheng L, Gong J, Yu T, Zou Y, Zhang M, Nie L, Chen X, Yue Q, Liu Y, Mao Q, et al. Diffuse Midline Gliomas With Histone H3 K27M Mutation in Adults and Children: A Retrospective Series of 164 Cases. Am J Surg Pathol. 2022;46(6):863–71. Jiang H, Yang K, Ren X, Cui Y, Li M, Lei Y, Lin S. Diffuse midline glioma with H3 K27M mutation: a comparison integrating the clinical, radiological, and molecular features between adult and pediatric patients. Neuro Oncol. 2020;22(5):e1–9. Maeda S, Ohka F, Okuno Y, Aoki K, Motomura K, Takeuchi K, Kusakari H, Yanagisawa N, Sato S, Yamaguchi J, et al. H3F3A mutant allele specific imbalance in an aggressive subtype of diffuse midline glioma, H3 K27M-mutant. Acta Neuropathol Commun. 2020;8(1):8. Pang H, Dang X, Ren Y, Zhuang D, Qiu T, Chen H, Zhang J, Ma N, Li G, Zhang J, et al. 3D-ASL perfusion correlates with VEGF expression and overall survival in glioma patients: Comparison of quantitative perfusion and pathology on accurate spatial location-matched basis. J Magn Reson Imaging. 2019;50(1):209–20. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterialversion.docx Supplementary Information Additional file 1: Table S1 MRI protocols of T1w, T2w, FLAIR and contrast-enhanced T1w. Table S2 Univariate logistic regression analysis for clinical and imaging parameters to differentiate H3K27 -altered DMGs from non-DMGs patients. Table S3 Univariate ROC analysis for clinical and imaging parameters to differentiate H3K27 -altered DMGs from non-DMGs patients. Table S4 Performance of five models for Predicting H3K27 -altered DMGs. Table S5 Performance of each imaging modality alone for Predicting H3K27 -altered DMGs. Table S6 Evaluation indicators for the five Cox models. Table S7 Univariate Cox analysis of clinical and imaging features and OS. Table S8 Age information for all patients. Table S9 Molecular information for all patients. Fig. S1 the flowchart of this study. Fig. S2 Patient Age Distribution Chart. Fig. S3 The Delong test of the five multivariable models. Fig. S4 The ROC curves for each imaging modality alone. Fig. S5 The scatter diagram between Ki-67 and APTw. Cite Share Download PDF Status: Published Journal Publication published 19 Nov, 2025 Read the published version in BMC Medicine → Version 1 posted Editorial decision: Revision requested 10 Sep, 2025 Reviews received at journal 13 Jun, 2025 Reviewers agreed at journal 13 Jun, 2025 Reviews received at journal 13 Jun, 2025 Reviewers agreed at journal 11 Jun, 2025 Reviewers invited by journal 08 Jun, 2025 Submission checks completed at journal 06 Jun, 2025 First submitted to journal 05 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5128033","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":501486081,"identity":"830d32f2-2aac-45b6-9046-3db1d657b5ba","order_by":0,"name":"Jun Qiu","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Qiu","suffix":""},{"id":501486085,"identity":"07fa1775-91e8-401e-838f-bf481d949275","order_by":1,"name":"Junjie Li","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junjie","middleName":"","lastName":"Li","suffix":""},{"id":501486087,"identity":"e93347b4-b76d-4fb1-b0bd-9f228e146fa3","order_by":2,"name":"Yunyun Duan","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yunyun","middleName":"","lastName":"Duan","suffix":""},{"id":501486088,"identity":"e5ce5da4-cac3-4e51-8f12-a08354c7ff00","order_by":3,"name":"Jun Sun","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Sun","suffix":""},{"id":501486089,"identity":"28426d45-e560-43e0-ac2c-9ec9180117f7","order_by":4,"name":"Yuna Li","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuna","middleName":"","lastName":"Li","suffix":""},{"id":501486090,"identity":"dd6dc9ce-a789-4b8e-9b86-9495d1d8c55d","order_by":5,"name":"Min Guo","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Guo","suffix":""},{"id":501486091,"identity":"d7d638ba-25b9-4180-a3bf-57527449d256","order_by":6,"name":"Minghao Wu","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Minghao","middleName":"","lastName":"Wu","suffix":""},{"id":501486092,"identity":"c7b8ca95-b922-419b-918a-f8e6c38b1515","order_by":7,"name":"Xiaolu Xu","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaolu","middleName":"","lastName":"Xu","suffix":""},{"id":501486093,"identity":"8d8c158a-50ce-469c-8533-ebe0130f24c1","order_by":8,"name":"Tiantian Hua","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Tiantian","middleName":"","lastName":"Hua","suffix":""},{"id":501486094,"identity":"6652b379-48bc-4e24-9f69-772d2a250e06","order_by":9,"name":"Yuwei Liu","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuwei","middleName":"","lastName":"Liu","suffix":""},{"id":501486095,"identity":"cd2ea015-bc06-45c2-ba07-d80331b5c2e1","order_by":10,"name":"Ying Jin","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ying","middleName":"","lastName":"Jin","suffix":""},{"id":501486098,"identity":"2a893fc3-b707-490d-8803-bbc7c2c11c3a","order_by":11,"name":"Xing Liu","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University,","correspondingAuthor":false,"prefix":"","firstName":"Xing","middleName":"","lastName":"Liu","suffix":""},{"id":501486099,"identity":"d6420685-3feb-4dd8-a506-ae42ad648edf","order_by":12,"name":"Liwei Zhang","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Liwei","middleName":"","lastName":"Zhang","suffix":""},{"id":501486103,"identity":"a89fea65-edcb-4951-8393-b6b293cca909","order_by":13,"name":"Zhizheng Zhuo","email":"","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhizheng","middleName":"","lastName":"Zhuo","suffix":""},{"id":501486105,"identity":"5570d727-a8c7-435f-8a97-ba9c5fedb4ec","order_by":14,"name":"Yaou Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyUlEQVRIiWNgGAWjYBACPmYILWfAcICBmSgtbFBlxgYMhxmbidMCpRM3MDATq4Wdx0ziR8Xh9O2M548/LmCwyZd3IOgwHjPJnjOHc3c2AB02gyHNcuMBIrRI8LYdzt1wAKiFh+GwgWEDMbb8bTucbkCSFmmgLQlwLfIEdAC1sBVby5xJNwT6xXA2j0GagQEhLfz8hzfefFNhLW8ucfDBZ54KGwN5Qg4DAhYJMCVxAEgArTA4QFgL8weIfVDTibFlFIyCUTAKRhYAAJGuOgHHNsLxAAAAAElFTkSuQmCC","orcid":"","institution":"Beijing Tiantan Hospital, Capital Medical University","correspondingAuthor":true,"prefix":"","firstName":"Yaou","middleName":"","lastName":"Liu","suffix":""}],"badges":[],"createdAt":"2024-09-21 09:59:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5128033/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5128033/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12916-025-04472-6","type":"published","date":"2025-11-19T15:56:53+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":90284044,"identity":"6e8ffa86-2517-4bd5-8c5c-7a2bff9a5854","added_by":"auto","created_at":"2025-09-01 05:47:22","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3388912,"visible":true,"origin":"","legend":"\u003cp\u003eAn overview of this study. T1w = T1-weighted; T2w = T2-weighted; T1C = contrast-enhanced T1-weighted; APTw = amide proton transfer-weighted; ASL = arterial spin labeling; DKI = diffusion kurtosis imaging; rCBF = relative cerebral blood flow; MD = mean diffusivity.\u003c/p\u003e","description":"","filename":"figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-5128033/v1/d602c456aac173807880d5cd.png"},{"id":90284046,"identity":"2b1edf17-981f-4e1c-aa96-ebadd69d0be4","added_by":"auto","created_at":"2025-09-01 05:47:22","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3995990,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Representative conventional MRI, metabolic, perfusion, and diffusion metric maps of patients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMG and non-DMG. (a) A 5-year-old female with \u003cem\u003eH3K27\u003c/em\u003e-altered DMG; (b) A 42-year-old male with non-DMG (Grade 4 astrocytoma). (B) Differences in clinical and MRI variables between \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and non-DMGs groups. T1w = T1-weighted; T2w = T2-weighted; T1C = contrast-enhanced T1-weighted; APTw = amide proton transfer-weighted; rCBF = relative cerebral blood flow; FA = fractional anisotropy; MD = mean diffusivity; MK = mean kurtosis; AD = axial diffusivity; RD = radial diffusivity; * p \u0026lt; 0.05.** p \u0026lt; 0.01.*** p \u0026lt; 0.001\u003c/p\u003e","description":"","filename":"figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-5128033/v1/4d1c0916e3d44048d69e4ec0.png"},{"id":90285232,"identity":"1b670a4d-1a31-4b34-921b-668e25b16221","added_by":"auto","created_at":"2025-09-01 05:55:22","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1002795,"visible":true,"origin":"","legend":"\u003cp\u003e(A) ROC curves demonstrating the diagnostic performance for H3K27-altered DMGs across five models; (B) Forest plot of the prognostic Model 5; (C) Kaplan-Meier analysis of prognostic value using the median value of APTw; (D) Kaplan-Meier analysis of prognostic value using the median value of rCBF; (E) Kaplan-Meier analysis of the prognostic value of low- and high-risk groups. ROC = receiver operating characteristic; KPS = Karnofsky Performance Scale; GTR = gross-total resection; STR = subtotal resection; PR = partial resection; APTw = amide proton transfer-weighted; rCBF = relative cerebral blood flow; MD = mean diffusivity; FA = fractional anisotropy; HR = Hazard ratios; CI = confidence interval. * indicate p \u0026lt; 0.05\u003c/p\u003e","description":"","filename":"figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-5128033/v1/64b9c2c4d1d07b1629fac462.png"},{"id":96649995,"identity":"87b0223c-6f2b-42a3-9f27-eac8c0befc5d","added_by":"auto","created_at":"2025-11-24 16:04:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9499733,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5128033/v1/edc2835e-09de-4de3-abe8-d22b20f67cd4.pdf"},{"id":90285231,"identity":"b1dc1958-ba9b-4e8a-81ee-f59e04173c46","added_by":"auto","created_at":"2025-09-01 05:55:22","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1911903,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Information\u003c/p\u003e\n\u003cp\u003eAdditional file 1: Table S1 MRI protocols of T1w, T2w, FLAIR and contrast-enhanced T1w. Table S2 Univariate logistic regression analysis for clinical and imaging parameters to differentiate \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs from non-DMGs patients. Table S3 Univariate ROC analysis for clinical and imaging parameters to differentiate \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs from non-DMGs patients. Table S4 Performance of five models for Predicting \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Table S5 Performance of each imaging modality alone for Predicting \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Table S6 Evaluation indicators for the five Cox models. Table S7 Univariate Cox analysis of clinical and imaging features and OS. Table S8 Age information for all patients. Table S9 Molecular information for all patients. Fig. S1 the flowchart of this study. Fig. S2 Patient Age Distribution Chart. Fig. S3 The Delong test of the five multivariable models. Fig. S4 The ROC curves for each imaging modality alone. Fig. S5 The scatter diagram between Ki-67 and APTw.\u003c/p\u003e","description":"","filename":"Supplementarymaterialversion.docx","url":"https://assets-eu.researchsquare.com/files/rs-5128033/v1/cf58f33d4dda0a2c508a654a.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Metabolic, Perfusion, and Diffusion Imaging Enhance Diagnosis and Prognosis of H3K27-Altered Diffuse Midline Gliomas","fulltext":[{"header":"Background","content":"\u003cp\u003e \u003cem\u003eH3K27\u003c/em\u003e-altered diffuse midline gliomas (DMGs) are fatal malignancies of the central nervous system (CNS). According to the 2021 World Health Organization (WHO) classification, \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs are classified as WHO Grade 4 tumors and characterized by diffuse and aggressive behavior[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e].DMGs are categorized as a pediatric-type diffuse high-grade glioma in the CNS WHO 2021[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs predominate in children but can also be seen in adolescents and adults[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. These tumors constitute approximately 80% of brainstem tumors[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] and primarily develop within midline structures[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Previous studies usually compared adult DMGs with midline \u003cem\u003eIDH\u003c/em\u003e wild-type glioblastomas (GBMs)[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. However, in clinical practice, the differential diagnosis of brainstem tumors often involves a wider variety of glioma and glioneuronal subtypes, such as astrocytomas[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], pilocytic astrocytomas[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e], and gangliogliomas[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], in addition to GBMs. They often present with overlapping clinical and imaging characteristics. Accurate differentiation of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs from \u003cem\u003eH3K27\u003c/em\u003e wide-type midline tumors (referred as non-DMGs) is crucial because of the significant variation in therapeutic strategies and prognostic implications.\u003c/p\u003e \u003cp\u003eMagnetic resonance imaging (MRI) is the standard non-invasive examination for diagnosing DMGs and monitoring tumor response to treatment. Conventional MRI, combined with radiomics and deep learning, has been utilized to predict \u003cem\u003eH3K27\u003c/em\u003e genetic alterations[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] and prognosis[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, these engineered features often lack interpretability, particularly in clinical practice. Furthermore, conventional MRI primarily provides anatomical insights into tumors and fails to capture the biological information related to \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs.\u003c/p\u003e \u003cp\u003eIn the past decade, advanced MRI techniques such as amide proton transfer-weighted (APTw), arterial spin labeling (ASL), and multi-shell diffusion imaging have been utilized to characterize metabolism[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], angiogenesis[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], and cellular density[\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] within brain tumors. Recently, several studies have applied these imaging techniques to explore the underlying pathology and aid in clinical diagnosis of DMGs. Zhuo et al.[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] demonstrated that APTw value and APTw image-derived radiomic features had good predictive ability for \u003cem\u003eH3K27\u003c/em\u003e mutation status, offering a novel imaging biomarker for the non-invasive assessment of DMGs. Studies also found that \u003cem\u003eH3K27\u003c/em\u003e-mutant tumors typically exhibit high cerebral blood flow (CBF), indicating increased angiogenic activity and invasiveness[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Furthermore, diffusion-weighted imaging (DWI) has been used to investigate the biological characteristics of DMGs, with lower apparent diffusion coefficient (ADC) values often associated with higher cellular density and poorer prognosis[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. However, DWI is based on a Gaussian distribution model and cannot fully capture cellular complexity. In contrast, multi-shell diffusion MRI, such as diffusion kurtosis imaging (DKI), provides non-Gaussian diffusion information, offering a more comprehensive view of tissue complexity and microenvironmental changes[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. A recent study using multi-shell diffusion MRI-derived metrics showed high discriminative power in distinguishing between \u003cem\u003eH3K27\u003c/em\u003e-altered and wild-type midline gliomas[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. However, most previous studies have focused on identifying DMGs using a single imaging technique, failing to fully leverage the integration of metabolic, perfusion, and diffusion imaging for diagnosing \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. In addition, no study has used these MRI techniques to predict prognosis of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, which is vital for treatment monitoring and clinical management.\u003c/p\u003e \u003cp\u003eTo better reflect this clinical reality, our study incorporated a broader spectrum of non-DMG midline tumors as the control group. This study aims to characterize the metabolic, perfusion, and diffusion characteristics of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs using APTw, ASL, and DKI. Furthermore, we systematically assess their added value for diagnosing \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and predicting prognosis, compared with conventional MRI.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design\u003c/h2\u003e \u003cp\u003eThe overview study design, including data collection, image processing, feature extraction, and statistical analysis, is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEthics Approval\u003c/h3\u003e\n\u003cp\u003e This study was approved by the Animal and Human Ethics Committee of Beijing Tiantan Hospital, Capital Medical University (KY2022-078-02). Written informed consent was obtained from all patients or their legal guardians in accordance with the Declaration of Helsinki.\u003c/p\u003e\n\u003ch3\u003eParticipants\u003c/h3\u003e\n\u003cp\u003eFrom June 2020 to May 2023, 125 patients with clinically and radiologically suspected brainstem tumors were prospectively recruited. Inclusion criteria were: (1) no prior treatment; (2) complete preoperative MRI examinations; (3) tumors were located in the brainstem, confirmed by MRI; and (4) scheduled for biopsy or surgical resection and with a final histopathological diagnosis. Exclusion criteria were: (1) poor image quality; and (2) incomplete histological information.\u003c/p\u003e \u003cp\u003eClinical information collected included age, sex, Karnofsky Performance Status (KPS) score at admission or clinic, and tumor histology. All patients underwent regular follow-up after diagnosis. Patients lost to follow-up were censored at their last known contact date. Overall survival (OS) was defined as the time from diagnosis to death from any cause. The final follow-up was on April 1, 2024. The median follow-up was 11.7 months (95% confidence interval [CI]: 3.7\u0026ndash;28.8 months).\u003c/p\u003e\n\u003ch3\u003ePathological Analysis\u003c/h3\u003e\n\u003cp\u003eThe tumors were classified by an experienced pathologist (X.L, with 10 years of experience) according to the 2021 WHO classification of CNS tumors[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. \u003cem\u003eH3K27\u003c/em\u003e alteration status was determined by immunohistochemical using a mutation-specific antibody (Merck Millipore, Billerica, MA, USA). The Ki-67 index was assessed by immunohistochemistry using the MIB-1 (anti-Ki-67) antibody (Dako, Denmark). Tumor cells were identified based on standard morphological criteria. In five hot spots (400\u0026times; magnification), the total number of tumor cells and Ki-67- positive cells were manually counted. The Ki-67 index was calculated as the mean percentage of positive cells across these fields.\u003c/p\u003e\n\u003ch3\u003eMRI Acquisition\u003c/h3\u003e\n\u003cp\u003eMRI acquisition was performed on a 3T MR scanner (Ingenia CX, Philips Healthcare, Best, The Netherlands) with a 32-channel head coil. The MR protocol included T1-weighted image(T1w), T2-weighted image(T2w), and contrast-enhanced T1-weighted imaging(T1C). Details of the conventional sequences can be found in Additional file 1: Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eFor APTw, the protocol parameters for 3D fast spin-echo sequences were: repetition time (TR)\u0026thinsp;=\u0026thinsp;5864 ms, echo time (TE)\u0026thinsp;=\u0026thinsp;8.8 ms, flip angle (FA)\u0026thinsp;=\u0026thinsp;90\u0026deg;, voxel size\u0026thinsp;=\u0026thinsp;2 mm \u0026times; 2 mm \u0026times; 6 mm, with 7 slices acquired. Saturation pulses at +\u0026thinsp;3.5 ppm (relative to the water resonance frequency) were applied with a B1 amplitude of 2 \u0026micro;T for 2 s to saturate amide protons. To normalize and interpolate the Z-spectrum, six additional volumes were acquired at offset frequencies (\u0026plusmn;\u0026thinsp;3.1 ppm, \u0026minus;\u0026thinsp;3.5 ppm, \u0026plusmn; 3.9 ppm, and \u0026minus;\u0026thinsp;1560 ppm). Three further acquisitions at +\u0026thinsp;3.5 ppm with varied echo-time shifts generated Dixon-type B0 field maps for correcting B0 inhomogeneity in the Z-spectrum frequency domain[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. SENSE factor\u0026thinsp;=\u0026thinsp;1.6, acquisition time\u0026thinsp;=\u0026thinsp;1 min 54 s.\u003c/p\u003e \u003cp\u003ePseudo-continuous arterial spin labeling (pCASL) was used for ASL data acquisition. The acquisition parameters were as follows: 3D gradient and spin-echo imaging, 20 slices, 8 pairs of control/label, TR/TE\u0026thinsp;=\u0026thinsp;3903 ms/11 ms, SENSE factor\u0026thinsp;=\u0026thinsp;1.3, labeling duration\u0026thinsp;=\u0026thinsp;1800 ms, post-labeling delay (PLD)\u0026thinsp;=\u0026thinsp;1500 ms, scan time\u0026thinsp;=\u0026thinsp;3 min 31 s.\u003c/p\u003e \u003cp\u003eDKI was acquired using axial 2D spin-echo echo-planar imaging (SE-EPI). Acquisition parameters were as follows: TR/TE\u0026thinsp;=\u0026thinsp;4000 ms/88 ms, FA\u0026thinsp;=\u0026thinsp;90\u0026deg;, in-plane voxel size\u0026thinsp;=\u0026thinsp;2.5 mm \u0026times; 2.5 mm, acquisition matrix size\u0026thinsp;=\u0026thinsp;96 \u0026times; 96, slice thickness\u0026thinsp;=\u0026thinsp;2.5 mm, number of slices\u0026thinsp;=\u0026thinsp;60, b-values\u0026thinsp;=\u0026thinsp;0, 1000, 2000 s/mm\u0026sup2;, with 48 diffusion-sensitizing gradient directions for each non-zero b-value, SENSE factor\u0026thinsp;=\u0026thinsp;2, multiband factor\u0026thinsp;=\u0026thinsp;3, acquisition time\u0026thinsp;=\u0026thinsp;6 min 48 s. Additional phase-encoding reverse B0 images were acquired for distortion correction.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eTumor Segmentation\u003c/h2\u003e \u003cp\u003eFor tumor segmentation, the solid tumor was manually delineated on T2w images, excluding cystic, necrotic, and hemorrhagic areas. Segmentation was performed by two neuroradiologists (M.W. and J.Q., with over 5 and 10 years of neuroradiology experience, respectively), who were blinded to the \u003cem\u003eH3K27\u003c/em\u003e alteration status. T1w and T1C images served as anatomical references. Segmentation was executed using 3D Slicer software (version 4.10.2; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"http://www.slicer.org\" target=\"_blank\"\u003ewww.slicer.org\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.slicer.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). All segmentations were subsequently reviewed and, if necessary, refined by a senior neuroradiologist (Y.D., with 20 years of experience). The Dice similarity coefficient between the two raters for the solid tumor segmentation was 0.89. The intersection of the two reviewers\u0026rsquo; segmentations was used as the final tumor mask. Tumor volume was calculated from this final segmentation.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eImage Processing\u003c/h3\u003e\n\u003cp\u003eAPTw images were automatically processed using the embedded post-processing pipeline on the MR console. Magnetization transfer ratio asymmetry (MTRasym) analysis was conducted voxel-by-voxel to differentiate APT signals from background effects[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. B0 field maps were used to correct for inhomogeneity artifacts, and the MTRasym value at +\u0026thinsp;3.5 ppm was calculated and expressed as a percentage. To extract APTw values, T2w images, and tumor segmentation were coregistered to the corresponding APTw images, and the mean APTw value within the solid tumor was measured for each case.\u003c/p\u003e \u003cp\u003eRelative cerebral blood flow (rCBF) was calculated from ASL data using the embedded post-processing pipeline. To account for inter-individual variation in perfusion, the rCBF maps were normalized by subtracting the mean gray-matter rCBF and dividing by that same mean value. Cerebral gray matter was segmented from T1w images using Statistical Parametric Mapping (SPM12, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.fil.ion.ucl.ac.uk/spm/\u003c/span\u003e\u003cspan address=\"https://www.fil.ion.ucl.ac.uk/spm/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and coregistered to the ASL source images. Next, T2w images and tumor segmentation mask were coregistered to the corresponding ASL source images. Finally, the mean normalized rCBF within the tumor mask was extracted for each case;\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:rCBF=\\frac{Lesional\\:CBF-Cerebellar\\:Hemispℎeric\\:CBF}{Cerebellar\\:Hemispℎeric\\:CBF}\\)\u003c/span\u003e\u003c/span\u003e༛ Normalized rCBF is referred to simply as rCBF in subsequent analyses.\u003c/p\u003e \u003cp\u003eDKI images were processed using DIPY (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dipy.org/\u003c/span\u003e\u003cspan address=\"https://dipy.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), FMRIB Software Library (FSL) version 6.0 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL\u003c/span\u003e\u003cspan address=\"https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and MRtrix3 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.mrtrix.org/\u003c/span\u003e\u003cspan address=\"https://www.mrtrix.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Preprocessing steps included denoising, distortion correction, eddy-current correction, and skull stripping. Seven diffusion metrics were calculated: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), mean kurtosis (MK), axial kurtosis (AK), and radial kurtosis (RK). To extract diffusion metrics, T2w images, and tumor segmentation mask were coregistered to the corresponding B0 images. Finally, the mean diffusion metrics within the solid tumor were measured for each patient.\u003c/p\u003e\n\u003ch3\u003eRadiological Evaluation\u003c/h3\u003e\n\u003cp\u003eTwo independent neuroradiologists (M.W. and J.Q., with over 5 and 10 years of neuroradiology experience, respectively) characteristics on conventional MRI sequences (T1w, T2w, and T1C), including: (1) tumor location (midbrain, pons, or medulla); (2) presence of hydrocephalus; (3) contrast enhancement patterns; and (4) necrotic. Hydrocephalus was diagnosed based on MRI findings of the lateral ventricles (Evans index\u0026thinsp;\u0026gt;\u0026thinsp;0.3)[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. The extent of resection (EOR) on postoperative images was classified as gross total resection (GTR: 100% EOR), subtotal resection (STR: \u0026gt; 50% and \u0026lt;\u0026thinsp;100% EOR), and partial resection (PR: \u0026lt; 50% EOR). Discrepancies were resolved by a senior neuroradiologist (Y.D. with over 20 years of neuroradiology experience), who reviewed and confirmed the final assessments.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eData analysis was performed using R statistical software (version 4.2.3). Categorical variables were reported as frequencies and percentages and analyzed by the Chi-square test. Continuous variables were compared using Student's t-test or the Mann-Whitney U test, depending on data distribution.\u003c/p\u003e \u003cp\u003eFirst, we assessed the separate and combined contributions of APTw, ASL, and DKI-derived measures to the diagnosis of DMGs. Multivariate logistic regression with leave-on-out cross-validation (LOOCV) was used to distinguish \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs from non-DMGs using five models (Model 1: clinical variables and conventional MR features; Model 2: Model 1\u0026thinsp;+\u0026thinsp;APTw; Model 3: Model 1\u0026thinsp;+\u0026thinsp;ASL; Model 4: Model 1\u0026thinsp;+\u0026thinsp;DKI; and Model 5: Model 1\u0026thinsp;+\u0026thinsp;APTw\u0026thinsp;+\u0026thinsp;ASL\u0026thinsp;+\u0026thinsp;DKI). We then calculated the area under the curve (AUC), sensitivity, specificity, and accuracy to evaluate model performance. Differences between models were assessed using DeLong\u0026rsquo;s test. Second, we assessed the separate and integrated contributions of APTw, ASL, and DKI metrics to the prognosis of DMGs. Multivariate Cox regression with LOOCV was used to build five prognostic models (similar as above five models in diagnostic tasks) to predict OS in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs group. Model performance was assessed by the concordance index (C-index). Additional survival analyses stratified by these metrics and risk predictions by Cox regression (Cox model 5) were conducted using the Kaplan-Meier method and log-rank test. A p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003eDemographic, Clinical, and Conventional MRI characteristics\u003c/h2\u003e\n \u003cp\u003eA total of 125 patients with brainstem glioma were initially eligible. Four patients were excluded for prior treatment before biopsy or surgery. Six for poor APTw image quality. Five for poor ASL image quality. Ten patients for poor DKI image quality and five for missing histological information (Additional file 1: Fig. \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eUltimately, 95 patients were included: 71 patients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and 24 patients with non-DMGs. The non-DMG group comprised seven pilocytic astrocytomas, two gangliogliomas, seven wild-type GBMs, six astrocytomas (including two Grade 4, one Grade 3, and three Grade 2), one mixed glioma-neuronal tumor, and one gliosis. Representative conventional MRI and quantitative parameter maps (APTw for metabolism, ASL-CBF for perfusion, DKI-derived MD for diffusion) from two representative cases are illustrated in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eA.\u003c/p\u003e\n \u003cp\u003ePatients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs were younger than those with non-DMGs (15.13\u0026thinsp;\u0026plusmn;\u0026thinsp;1.58 years vs. 22.46\u0026thinsp;\u0026plusmn;\u0026thinsp;3.60 years, p\u0026thinsp;=\u0026thinsp;0.035); the age distribution of patients is shown in Additional file 1: Fig. \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e. Mean KPS was lower in \u003cem\u003eH3K27\u003c/em\u003e-altered DMG group than in non-DMGs group (86.06 vs. 97.08, p\u0026thinsp;=\u0026thinsp;0.012). Most \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs were located in the pons (66/71; 92.96%). Non-DMGs were located in the pons (62.50%), medulla (8.33%), and midbrain (29.17%). Significant differences in EOR were observed between the H3K27-altered DMG and non-DMG groups: biopsy (40.85% vs. 16.67%); GTR (16.90% vs. 62.50%); STR (9.86% vs. 16.67%); PR (32.39% vs. 4.16%) (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). There were no differences between \u003cem\u003eH3K27\u003c/em\u003e-altered DMG and non-DMG groups in conventional MRI features (e.g., enhancement, necrosis, hydrocephalus) (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDemographics, clinical, and conventional MRI characteristics of patients with DMGs and non-DMGs\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAll Patients (\u003cem\u003eN\u0026thinsp;=\u0026thinsp;95\u003c/em\u003e)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eH3K27\u003c/em\u003e-altered DMGs (\u003cem\u003eN\u0026thinsp;=\u0026thinsp;71\u003c/em\u003e)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003enon-DMGs (\u003cem\u003eN\u0026thinsp;=\u0026thinsp;24\u003c/em\u003e)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep-value\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16.98\u0026thinsp;\u0026plusmn;\u0026thinsp;1.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15.13\u0026thinsp;\u0026plusmn;\u0026thinsp;1.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.46\u0026thinsp;\u0026plusmn;\u0026thinsp;3.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.035*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSex (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003emale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e37 (38.95% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e28 (39.44% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (37.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003efemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e58 (61.05% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e43 (60.56% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (62.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEOR (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBiopsy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e33 (34.74% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e29 (40.85% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (16.67% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGTR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e27 (28.42% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12 (16.90% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (62.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSTR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (11.58% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (9.86% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (16.67% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e24 (25.26% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (32.39% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (4.16% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eKPS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e88.84\u0026thinsp;\u0026plusmn;\u0026thinsp;1.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e86.06\u0026thinsp;\u0026plusmn;\u0026thinsp;2.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e97.08\u0026thinsp;\u0026plusmn;\u0026thinsp;1.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.012*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLocation (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePons\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e81 (85.26% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e66 (92.96% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (62.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMedulla\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (7.37% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 (7.04% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 (8.33% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMidbrain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (7.37% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0 (0.00% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (29.17% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEnhancement (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eno\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e35 (36.84% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e26 (36.62% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (37.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eyes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e60 (63.16% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e45 (63.38% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (62.50% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNecrosis (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.767\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eno\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e48 (50.53% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e37 (52.11% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (45.83% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eyes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e47 (49.47% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e34 (47.89% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (54.17% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHydrocephalus (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eno\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e65 (68.42% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e49 (69.01% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16 (66.67% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eyes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30 (31.58% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22 (30.99% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (33.33% )\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVolume (cm\u003csup\u003e3\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.7 [15.3;29.7]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23.4 [19.0;30.0]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16.4 [10.4;24.5]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.008*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAPTw\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.74 [2.07;3.45]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.79 [2.30;3.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.01 [1.63;2.74]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erCBF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.35 [-0.50;-0.18]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.30 [-0.46;-0.17]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-0.49 [-0.61;-0.34]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.23 [0.20;0.27]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.24 [0.20;0.28]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.20 [0.18;0.25]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.015*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMD (\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003emm/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.34 [1.09;1.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.24 [1.02;1.49]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.43 [1.34;1.78]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAD (\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003emm/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.65 [1.38;1.91]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.57 [1.32;1.83]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.78 [1.65;2.05]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRD (\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003emm/s)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.19 [0.96;1.44]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.08 [0.86;1.35]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.30 [1.18;1.63]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMK\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.58 [0.51;0.68]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.58 [0.52;0.67]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.57 [0.46;0.72]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.837\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAK\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.57 [0.49;0.63]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.57 [0.50;0.62]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.54 [0.46;0.66]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.461\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRK\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.61 [0.54;0.75]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.61 [0.55;0.73]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.64 [0.47;0.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.932\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eEOR\u0026thinsp;=\u0026thinsp;extent of resection; GTR\u0026thinsp;=\u0026thinsp;gross-total resection; STR\u0026thinsp;=\u0026thinsp;subtotal resection, PR\u0026thinsp;=\u0026thinsp;partial resection; KPS\u0026thinsp;=\u0026thinsp;Karnofsky Performance Scale; APTw\u0026thinsp;=\u0026thinsp;amide proton transfer-weighted; rCBF\u0026thinsp;=\u0026thinsp;relative cerebral blood flow; FA\u0026thinsp;=\u0026thinsp;fractional anisotropy; MD\u0026thinsp;=\u0026thinsp;mean diffusivity; RD\u0026thinsp;=\u0026thinsp;radial diffusivity; AD\u0026thinsp;=\u0026thinsp;axial diffusivity; AK\u0026thinsp;=\u0026thinsp;axial kurtosis; MK\u0026thinsp;=\u0026thinsp;mean kurtosis; RK\u0026thinsp;=\u0026thinsp;radial kurtosis; HR\u0026thinsp;=\u0026thinsp;Hazard ratio. * indicate p\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003eMetabolic, Perfusion, and Diffusion Characteristics of DMGs\u003c/h2\u003e\n \u003cp\u003eFor metabolic, perfusion, and diffusion metrics, patients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs had significantly different values compared with non-DMG patients for APTw (median 2.79 vs. 2.01); rCBF (median \u0026minus;\u0026thinsp;0.30 vs. -0.49); FA (median 0.24 vs. 0.20); MD (median 1.24 vs. 1.43); AD (median 1.57 vs. 1.78); and RD (median 1.08 vs. 1.30) (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). However, DKI metrics (MK, AK, and RK) did not show significant differences (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eB).\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eMetabolic, Perfusion, and Diffusion Imaging Improve Diagnostic Performance for\u003c/strong\u003e \u003cstrong\u003eH3K27\u003c/strong\u003e\u003cstrong\u003e-Altered DMGs\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eWe Constructed five multivariate logistic regression models using variables that were significant in univariate analyses. Univariate logistic regression analysis results are shown in Additional file 1: Table \u003cspan class=\"InternalRef\"\u003eS2\u003c/span\u003e. Due to high collinearity among diffusion metrics. Because MD equals the average of AD and RD, it serves as a composite metric reflecting both diffusion components[\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e]. So, MD was selected for multivariable logistic and Cox regression analyses (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eFive multivariable regression models to identify H3K27-altered DMGs from non-DMGs\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 2\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 3\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 4\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 5\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eOR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eOR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eOR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eOR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eOR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLocation\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePons\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMedulla\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.23(0.32,22.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.447\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.19(0.24,26.99)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.505\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.53(0.31,29.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.414\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.53(0.17,20.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.722\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.41(0.11,23.26)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.798\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMidbrain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.97(0.93,1.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.183\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.98(0.94,1.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.341\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.96(0.96,1.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.079\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.95(0.90,0.99)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.029*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.95(0.89,1.00)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.043*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eKPS\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.91(0.81,0.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.058\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.92(0.81,0.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.064\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.92(0.81,0.99)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.099\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.90(0.77,0.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.066\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.94(0.82,1.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.203\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eVolume\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.03(0.97,1.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.411\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.96,1.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.614\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.94,1.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.991\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.05(0.99,1.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.154\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.95,1.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.578\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAPTw\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e3.81(1.65,10.65)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.004*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e3.41(1.39,10.28)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.014*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003erCBF\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e7.74(3.18,44.25)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.019*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.26(0.25,18.88)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.265\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eMD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.03(0.001,0.38)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.012*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.05(0.002,0.80)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.049*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eFA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.88(0.16,4.57)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.877\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.98(0.14,6.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.982\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"11\"\u003eKPS\u0026thinsp;=\u0026thinsp;Karnofsky Performance Scale; APTw\u0026thinsp;=\u0026thinsp;amide proton transfer-weighted; rCBF\u0026thinsp;=\u0026thinsp;relative cerebral blood flow ;FA\u0026thinsp;=\u0026thinsp;fractional anisotropy; MD\u0026thinsp;=\u0026thinsp;mean diffusivity; OR\u0026thinsp;=\u0026thinsp;odds ratios; CI\u0026thinsp;=\u0026thinsp;confidence interval; ref\u0026thinsp;=\u0026thinsp;reference; NA, not available. * indicates p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eModel 1 included clinical variables and conventional MRI features.\u003c/p\u003e\n \u003cp\u003eModel 2 added APTw measures to Model 1 and achieved an odds ratio (OR) of 3.81 (95% CI: 1.65\u0026ndash;10.65, p\u0026thinsp;=\u0026thinsp;0.004) with an AUC was 0.901, a 0.07 improvement over Model 1 (Delong\u0026rsquo;s test p\u0026thinsp;=\u0026thinsp;0.125).\u003c/p\u003e\n \u003cp\u003eModel 3 added rCBF measures to Model 1 and achieved an OR of 7.74 (95% CI: 3.18\u0026ndash;44.25, p\u0026thinsp;=\u0026thinsp;0.019) with an AUC was 0.881, a 0.05 improvement over Model 1 (Delong\u0026rsquo;s test p\u0026thinsp;=\u0026thinsp;0.216).\u003c/p\u003e\n \u003cp\u003eModel 4 added DKI metrics (MD and FA) to Model 1, MD remained an independent predictor (an OR of 0.03; 95% CI: 0.001\u0026ndash;0.38, p\u0026thinsp;=\u0026thinsp;0.012), and an AUC of Model was 0.904, a 0.073 improvement over Model 1 (Delong\u0026rsquo;s test p\u0026thinsp;=\u0026thinsp;0.046).\u003c/p\u003e\n \u003cp\u003eModel 5 integrated APTw, rCBF, MD, and FA into Model 1, APTw and MD remained significant independent contributors, with ORs of 3.41 (95% CI: 1.39\u0026ndash;10.28, p\u0026thinsp;=\u0026thinsp;0.014) and 0.05 (95% CI: 0.002- 0.80, p\u0026thinsp;=\u0026thinsp;0.049), respectively. The AUC for Model 5 was 0.994, a 0.163 improvement over Model 1 (Delong\u0026rsquo;s test p\u0026thinsp;=\u0026thinsp;0.030).\u003c/p\u003e\n \u003cp\u003eModel 5 (clinical variables\u0026thinsp;+\u0026thinsp;conventional MRI\u0026thinsp;+\u0026thinsp;APTw\u0026thinsp;+\u0026thinsp;ASL\u0026thinsp;+\u0026thinsp;DKI) achieved the highest AUC among all models, indicating the best diagnostic performance (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eA). The AUC and cutoff values for the diagnostic prediction of each variable are presented in Additional file 1: Table S3. Performance of five models for the diagnostic prediction in Additional file 1: Table S4. Delong\u0026rsquo;s test result is in Additional file 1: Fig. S3. The AUC values of the ROC curves obtained by using each imaging technique alone (APTw, ASL, DKI) have been provided in Additional file 1: Fig. S4 and Additional file 1: Table S5.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eMetabolic, Perfusion, and Diffusion Imaging Improve Prognostic Prediction for\u003c/strong\u003e \u003cstrong\u003eH3K27\u003c/strong\u003e\u003cstrong\u003e-Altered DMGs\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eAmong the 71 DMG patients with available OS data, we constructed five multivariate Cox regression models mirroring the diagnostic analysis framework (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). Variables achieving statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) on univariate COX regression analyses were selected for multivariable analysis.\u0026nbsp;\u003c/p\u003e\n \u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eMultivariable Cox regression models for predicting OS in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 2\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 3\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 4\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eModel 5\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eHR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eHR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eHR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eHR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eHR (95% CI)\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLocation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePons\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMedulla\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3.99(0.90, 17.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.068\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3.56(0.79, 16.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4.87(1.09, 21.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6.25(1.10, 35.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4.78(0.77, 29.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.093\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.98, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.876\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.97, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.737\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.97, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.805\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.97, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.788\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.99(0.97, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.630\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eKPS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.98, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.802\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.99, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.619\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.01(0.99, 1.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.583\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00(0.98, 1.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.932\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.01(0.99, 1.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.521\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEOR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBiopsy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eref\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGTR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.28(0.08, 0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.038*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.21(0.06, 0.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.012*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.27(0.08, 0.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.033*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.23(0.06, 0.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.023*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.20(0.05, 0.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.014*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSTR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.52(0.15, 1.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.286\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.45(0.13, 1.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.230\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.51(0.15, 1.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.332\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.64(0.18, 2.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.482\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.49(0.13, 1.86)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.294\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.92(0.43, 1.97)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.781\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.72(0.33, 1.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.386\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.79(0.35, 1.77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.615\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.96(0.45, 2.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.876\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.72(0.32, 1.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.428\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eKi-67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.04(1.02, 1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;\u0026thinsp;0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.03(1.01, 1.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.015*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.05(1.03, 1.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;\u0026thinsp;0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.04(1.02, 1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;\u0026thinsp;0.001*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.04(1.01, 1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.003*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVolume\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.03(1.00, 1.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.99, 1.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.185\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.99, 1.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.173\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.99, 1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02(0.99, 1.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.257\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAPTw\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.89(1.24, 2.87)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.003*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.66(1.03, 2.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.037*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erCBF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.87(1.46, 5.65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.15(1.05, 4.44)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.038*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.69(0.45, 6.34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.382\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.15(0.25, 5.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.856\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.72(0.62, 4.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.336\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.21(0.40, 3.69)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.733\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"11\"\u003eOS\u0026thinsp;=\u0026thinsp;overall survival; KPS\u0026thinsp;=\u0026thinsp;Karnofsky Performance Scale; EOR\u0026thinsp;=\u0026thinsp;extent of resection; GTR\u0026thinsp;=\u0026thinsp;gross-total resection; STR\u0026thinsp;=\u0026thinsp;subtotal resection, PR\u0026thinsp;=\u0026thinsp;partial resection; APTw\u0026thinsp;=\u0026thinsp;amide proton transfer-weighted; rCBF\u0026thinsp;=\u0026thinsp;relative cerebral blood flow; FA\u0026thinsp;=\u0026thinsp;fractional anisotropy; MD\u0026thinsp;=\u0026thinsp;mean diffusivity; HR\u0026thinsp;=\u0026thinsp;Hazard ratios; CI\u0026thinsp;=\u0026thinsp;confidence interval; ref\u0026thinsp;=\u0026thinsp;reference. * indicate p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n \u003cp\u003eModel 1 included clinical variables and conventional MRI features.\u003c/p\u003e\n \u003cp\u003eModel 2 added APTw to Model 1, APTw remained an independent predictor (hazard ratio [HR] of 1.89; 95% CI: 1.24\u0026ndash;2.87; p\u0026thinsp;=\u0026thinsp;0.003). The C-index of Model 2 was 0.714, a 0.009 decrease from Model 1 (C-index\u0026thinsp;=\u0026thinsp;0.723).\u003c/p\u003e\n \u003cp\u003eModel 3 added rCBF to Model 1, it yielded an HR of 2.87 (95% CI: 1.46\u0026ndash;5.65, p\u0026thinsp;=\u0026thinsp;0.002). The C-index of Model 3 was 0.745, a 0.022 improvement over Model 1.\u003c/p\u003e\n \u003cp\u003eModel 4 incorporating DKI metrics (MD and FA), an HR of MD was 1.69 (95% CI: 0.45\u0026ndash;6.34, p\u0026thinsp;=\u0026thinsp;0.382). The HR of FA was 1.72 (95% CI: 0.62\u0026ndash;4.80, p\u0026thinsp;=\u0026thinsp;0.336). The C-index of Model 4 was 0.729, a 0.006 increase over Model 1, but neither MD nor FA reached statistical significance.\u003c/p\u003e\n \u003cp\u003eModel 5 integrated APTw, rCBF, MD, and FA, APTw remained an independent predictor with an HR of 1.66 (95% CI: 1.03\u0026ndash;2.67, p\u0026thinsp;=\u0026thinsp;0.037), and rCBF also maintained its significance with an HR of 2.15 (95% CI: 1.05\u0026ndash;4.44, p\u0026thinsp;=\u0026thinsp;0.038). The C-index of Model 5 was 0.742, an improvement over Model 1.\u003c/p\u003e\n \u003cp\u003eAdditionally, Ki-67 index and GTR were significant prognostic factors in all models (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). Evaluation metrics for the five models are provided in Additional file 1: Table S6; univariate Cox results in Additional file 1: Table S7; Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eB shows the forest plot of Model 5. We further stratified patients by median APTw and rCBF values, log-rank p-values were 0.002 and 0.010, with HR values of 2.55 and 2.13, respectively (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eC, \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD). Predicted risks from Model 5 classified patients into high-risk and low-risk groups. The Kaplan-Meier curve for Model 5 showed an HR value of 8.57 (p\u0026thinsp;=\u0026thinsp;0.003) between the risk groups (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eE).\u003c/p\u003e\n \u003cp\u003eTo explore histological association of APTw, rCBF, and MD significantly that enhanced the diagnostic and prognostic performances of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, we correlated APTw, rCBF, and MD with the Ki-67 index. The results showed a significant correlation between Ki-67 and APTw (Additional file 1: Fig. S5), whereas rCBF and MD did not.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we used metabolic, perfusion, and diffusion imaging to assess the diagnostic performance and prognostic value of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. The main findings were as follows: (1) Patients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs were significantly younger and had lower KPS scores, and they showed significant differences in APTw, rCBF, FA, MD, AD, and RD values compared with non-DMGs. (2) APTw, rCBF, and MD each had independent predictive value for diagnosing \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, and the combination of APTw, rCBF, and MD (Model 5) achieved the highest diagnostic performance, with an AUC of 0.944, a 0.163 improvement compared with conventional MRI. (3) Both APTw and rCBF were independent prognostic factors for OS in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, and the combination of APTw, rCBF, MD, and clinical variables provided the most robust prognostic performance.\u003c/p\u003e \u003cp\u003e \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs are pediatric-type diffuse high-grade gliomas defined in the 2021 WHO classification[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. We found that patients with \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs were significantly younger than those without DMGs, reflecting tumor development in the developing brain[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. These patients also had lower KPS scores. Increased APTw values likely reflect active tumor cell proliferation[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Compared with non-DMGs, \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs exhibited higher rCBF values, which may relate to a positive correlation between rCBV and \u003cem\u003eH3K27M\u003c/em\u003e-positive nuclear density in tumor samples[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs also showed increased FA values and reduced diffusion metrics (MD, AD, and RD) values. FA measures diffusion anisotropy, indicating greater tissue complexity, while diffusion metrics (MD, AD, and RD) assess directional diffusion, reflecting restricted water movement in tissue[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. These findings suggest higher tumor cell density and more complex tissue microstructure in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs.\u003c/p\u003e \u003cp\u003eIn diagnostic models of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, APTw imaging assesses heterogeneous protein and peptide metabolism, which may reflect the histopathology and genetic alterations of gliomas[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Higher APTw values indicate increased protein and peptide metabolism and likely reflect active tumor cell proliferation[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. In our study, adding APTw values to clinical and conventional imaging features in the multivariable logistic regression model significantly improved diagnostic performance for \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. This finding aligns with Zhuo et al.[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], who showed that APTw-derived radiomics features predicted \u003cem\u003eH3K27\u003c/em\u003e mutations in brainstem glioma patients with 86% accuracy. Tumor perfusion also contributed to diagnostic accuracy for \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Our study showed that rCBF had significant diagnostic power in the multivariable model and independently predicted \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. This finding agrees with Kathrani et al.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], who found that rCBV ratios independently correlate with histone mutation status in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. We found that diffusion metrics (MD, AD, and RD) significantly predicted \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and enhanced diagnostic performance, potentially reflecting increased tumor cell density\u003csup\u003e16\u003c/sup\u003e. Xu et al.[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] made similar findings, suggesting that the densely packed tumor cells water diffusion, resulting in lower diffusion metrics values. This reduction may relate to tumor heterogeneity, including cellular proliferation and microenvironment changes[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. The combination of APTw, rCBF, and MD achieved the highest diagnostic performance for \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Integrating these metrics with clinical and conventional imaging features yielded an AUC of 0.944, a 0.163 improvement compared with conventional MRI, demonstrating the added diagnostic value of combining these metabolic, perfusion, and diffusion parameters.\u003c/p\u003e \u003cp\u003eIn the models predicting OS of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, APTw independently predicted OS, and patients with high APTw values had significantly lower OS (p\u0026thinsp;=\u0026thinsp;0.002). Supporting these findings, Joo et al.[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] also found that higher APT signals correlate with poorer prognosis in high-grade glioma patients. Previous studies suggest that higher APT signals reflect increased tumor cell proliferation and invasiveness and are associated with poor prognosis[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Our results confirm a certain significant correlation between the APTw values and the Ki-67 index, with a high Ki-67 index indicating rapid tumor cell proliferation[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. In \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, the Ki-67 index (Ki-67\u0026thinsp;\u0026le;\u0026thinsp;5% vs. \u0026gt;5%) showed prognostic value[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e], and several studies report differences in Ki-67 proliferation among DMG patients that correlate with biological behavior and prognosis[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Therefore, we investigated correlations between Ki-67 and imaging parameters to assess its potential as an auxiliary prognostic indicator. The results showed that rCBF also had significant prognostic ability and was an independent predictor of OS in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, with high rCBF values associated with shorter OS. Pang et al.[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e] also found that high CBF values are associated with poor prognosis in glioma patients. They observed that CBF correlates with \u003cem\u003eVEGF\u003c/em\u003e expression, suggesting that CBF may approximate angiogenesis across glioma grades. We observed that although diffusion metrics (MD, AD, and RD) were valuable for the diagnosis and classification of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, they did not significantly contribute to OS prediction, indicating that they may not fully cover factors influencing prognosis. Consequently, combining APTw, rCBF, and MD, along with clinical variables yielded the most robust prognostic performance by leveraging their complementary strengths. Furthermore, APTw and rCBF remained the primary independent predictors of OS, underscoring their value in prognostication for \u003cem\u003eH3K27\u003c/em\u003e-altered DMG patients.\u003c/p\u003e \u003cp\u003eOur results demonstrate the value of metabolic, perfusion, and diffusion imaging for diagnosing and predicting prognosis in \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Nevertheless, this study has certain limitations. First, our cohort included 95 patients (71 \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and 24 non-DMGs), resulting in class imbalance and a relatively small sample size. Although we used LOOCV for internal validation, this does not fully eliminate bias from sample size and class imbalance. Future studies should verify these results in larger, multicenter cohorts to improve the generalizability. Second, incomplete records of treatments beyond EOR prevented adjustment for all therapeutic confounders, potentially biasing the results; larger, multicenter cohorts with detailed treatment data are needed to address this issue. Third, the median follow-up of 11.7 months may be insufficient to assess long-term survival; future studies should include longer follow-up to verify the stability of these findings. Finally, we did not perform histopathological or longitudinal imaging validation to link imaging metrics with tumor biology over time. Future work should integrate histology, repeated imaging, or biopsy data within long-term cohorts to assess the biological significance and temporal stability of these biomarkers.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study improved diagnostic accuracy and prognostic prediction of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs by integrating metabolic, perfusion, and diffusion metrics with conventional MRI and clinical variables. The results demonstrate that integrating metabolic, perfusion, and diffusion imaging with conventional MRI and clinical variables provides robust support for accurate diagnosis and prognostic assessment of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eaxial diffusivity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eADC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eapparent diffusion coefficient\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAK\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eaxial kurtosis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAPTw\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eamide proton transfer-weighted\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eASL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003earterial spin labeling\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003earea under the curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDMGs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ediffuse midline gliomas\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003enon-DMGs\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003enon-\u003cem\u003eH3K27\u003c/em\u003e-altered DMGs\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCBF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecerebral blood flow\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003econfidence interval\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCNS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecentral nervous system\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDKI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ediffusion kurtosis imaging\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDWI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ediffusion-weighted imaging\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEOR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eextent of resection\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGTR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egross-total resection\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003efractional anisotropy\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHazard ratio\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKPS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eKarnofsky Performance Scale\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLOOCV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eleave-on-out-cross-validation\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emean diffusivity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMK\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emean kurtosis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMRI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMagnetic resonance imaging\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMTRasym\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMagnetization transfer ratio asymmetry\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eOR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eodds ratios\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eOS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eoverall survival\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003epCASL\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epseudo-continuous Arterial Spin Labeling\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePLD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epost-labeling delay\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epartial resection\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003erCBF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003erelative cerebral blood flow\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradial diffusivity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRK\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eradial kurtosis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ereceiver operating characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSE-EPI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003espin-echo echo-planar imaging\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSTR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003esubtotal resection\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eT1w\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eT1-weighted\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eT1C\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003econtrast-enhanced T1-weighted\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eT2w\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eT2-weighted\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eecho time\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003erepetition time\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eWHO\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eWorld Health Organization (WHO)\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Animal and Human Ethics Committee of Beijing Tiantan Hospital, Capital Medical University (KY2022-078-02). Written informed consent was obtained from all patients or their legal guardians in accordance with the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors approved the submitted version of the paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe corresponding authors had full control of the study data, and all data generated or analyzed during the study are available from the corresponding author by request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by Beijing Postdoctoral Foundation (NO.2-1-2-006-46); Beijing Natural Science Foundation (Grant Nos. 7252040, 7244328); Capital Health Development Scientific Research Special Project of Beijing Municipal Health Commission (NO.2022-1-2042); the Nursery engineering project of Beijing Tiantan Hospital Affiliated to Capital Medical University (No. 2023MP02); Capital Medical University (CCMU2024ZKYXZ007).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJun Qiu:\u003c/strong\u003e Conceptualization, Formal analysis, Investigation and Writing - Original Draft. \u003cstrong\u003eJunjie Li:\u0026nbsp;\u003c/strong\u003eVisualization,\u0026nbsp;Investigation. \u003cstrong\u003eYunyun Duan\u003c/strong\u003e\u003cstrong\u003e:\u0026nbsp;\u003c/strong\u003eInvestigation and Writing - Review \u0026amp; Editing.\u0026nbsp;\u0026nbsp;\u003cstrong\u003eJun Sun\u003c/strong\u003e\u003cstrong\u003e:\u003c/strong\u003e Software and Data Curation;\u0026nbsp;\u003cstrong\u003eYuna Li\u003c/strong\u003e: Resources and Formal analysis.\u0026nbsp;\u003cstrong\u003eMin Guo\u003c/strong\u003e\u003cstrong\u003e:\u003c/strong\u003e Resources.\u0026nbsp;\u003cstrong\u003eMinghao Wu\u003c/strong\u003e\u003cstrong\u003e:\u003c/strong\u003e Resources; \u003cstrong\u003eXiaolu Xu\u003c/strong\u003e\u003cstrong\u003e:\u003c/strong\u003e Investigation.\u0026nbsp;\u003cstrong\u003eTiantian Hua\u003c/strong\u003e:\u0026nbsp;Resources; \u003cstrong\u003eZ\u003c/strong\u003e\u003cstrong\u003ehizheng Zhuo:\u003c/strong\u003e\u0026nbsp; Project administration;\u0026nbsp;\u003cstrong\u003eYuwei Liu:\u003c/strong\u003e Supervision;\u0026nbsp;\u003cstrong\u003eYing Jin:\u003c/strong\u003e interpreted the data; \u003cstrong\u003eXing Liu:\u003c/strong\u003e technical support;\u003cstrong\u003e\u0026nbsp;Liwei Zhang:\u0026nbsp;\u003c/strong\u003eStudy design; \u003cstrong\u003eZhizheng Zhuo:\u003c/strong\u003e Supervision and Project administration,\u0026nbsp;Writing - Review \u0026amp; Editing. \u003cstrong\u003eYaou Liu:\u003c/strong\u003e Conceptualization, Supervision, Project administration, Funding acquisition, and Review \u0026amp; Editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' information:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e*\u003c/strong\u003eJun Qiu\u0026nbsp;\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui, 230036, China\u003c/p\u003e\n\u003cp\u003eJunjie Li\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eYunyun Duan\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eJun Sun\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eYuna Li\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eMin Guo\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eMinghao Wu\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eXiaolu Xu\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eTiantian Hua\u003c/p\u003e\n\u003cp\[email protected]\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eYuwei Liu\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eYing Jin\u003c/p\u003e\n\u003cp\[email protected]\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003eXing Liu\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Pathology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e#\u003c/strong\u003eLiwei Zhang\u003c/p\u003e\n\u003cp\u003eLiwei Zhang [email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing 10070, China\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e#\u003c/strong\u003eZhizheng Zhuo\u0026nbsp;\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e#\u003c/strong\u003eYaou Liu (Corresponding author)\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e\n\u003cp\u003e010-83911069\u003c/p\u003e\n\u003cp\u003eDepartment of Radiology, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, People's Republic of China\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLouis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison DW. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol. 2016;131(6):803\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLouis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D, Hawkins C, Ng HK, Pfister SM, Reifenberger G, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol. 2021;23(8):1231\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHayashi N, Fukai J, Nakatogawa H, Kawaji H, Yoshioka E, Kodama Y, Nakajo K, Uda T, Naito K, Kijima N, et al. Neuroradiological, genetic and clinical characteristics of histone H3 K27-mutant diffuse midline gliomas in the Kansai Molecular Diagnosis Network for CNS Tumors (Kansai Network): multicenter retrospective cohort. Acta Neuropathol Commun. 2024;12(1):120.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeach JL, Roebker J, Schafer A, Baugh J, Chaney B, Fuller C, Fouladi M, Lane A, Doughman R, Drissi R, et al. MR imaging features of diffuse intrinsic pontine glioma and relationship to overall survival: report from the International DIPG Registry. Neuro Oncol. 2020;22(11):1647\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLovibond S, Gewirtz AN, Pasquini L, Krebs S, Graham MS. The promise of metabolic imaging in diffuse midline glioma. Neoplasia. 2023;39:100896.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRaab P, Banan R, Akbarian A, Esmaeilzadeh M, Samii M, Samii A, Bertalanffy H, Lehmann U, Krauss JK, Lanfermann H et al. Differences in the MRI Signature and ADC Values of Diffuse Midline Gliomas with H3 K27M Mutation Compared to Midline Glioblastomas. Cancers (Basel) 2022, 14(6).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSim Y, Choi SH, Lee N, Park YW, Ahn SS, Chang JH, Kim SH, Lee SK. Clinical, qualitative imaging biomarkers, and tumor oxygenation imaging biomarkers for differentiation of midline-located IDH wild-type glioblastomas and H3 K27-altered diffuse midline gliomas in adults. Eur J Radiol. 2024;173:111384.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen LH, Pan C, Diplas BH, Xu C, Hansen LJ, Wu Y, Chen X, Geng Y, Sun T, Sun Y, et al. The integrated genomic and epigenomic landscape of brainstem glioma. Nat Commun. 2020;11(1):3077.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePan C, Li T, Zhang M, Wang Y, Xiao X, Zhang P, Zuo P, Wang Y, Xiao D, Wu Z, et al. A novel radiological classification of midbrain pilocytic astrocytomas and its implication for surgical management: a single-institution experience of 57 cases. J Neurosurg. 2024;140(6):1527\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAcharya S, Liu JF, Tatevossian RG, Chiang J, Qaddoumi I, Gajjar A, Walker D, Harreld JH, Merchant TE, Ellison DW. Risk stratification in pediatric low-grade glioma and glioneuronal tumor treated with radiation therapy: an integrated clinicopathologic and molecular analysis. Neuro Oncol. 2020;22(8):1203\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChauhan RS, Kulanthaivelu K, Kathrani N, Kotwal A, Bhat MD, Saini J, Prasad C, Chakrabarti D, Santosh V, Uppar AM, et al. Prediction of H3K27M mutation status of diffuse midline gliomas using MRI features. J Neuroimaging. 2021;31(6):1201\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePan CC, Liu J, Tang J, Chen X, Chen F, Wu YL, Geng YB, Xu C, Zhang X, Wu Z, et al. A machine learning-based prediction model of H3K27M mutations in brainstem gliomas using conventional MRI and clinical features. Radiother Oncol. 2019;130:172\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu X, Jiang Z, Roth HR, Anwar SM, Bonner ER, Mahtabfar A, Packer RJ, Kazerooni AF, Bornhorst M, Linguraru MG. Early prognostication of overall survival for pediatric diffuse midline gliomas using MRI radiomics and machine learning: A two-center study. Neurooncol Adv. 2024;6(1):vdae108.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evon Knebel Doeberitz N, Kroh F, Breitling J, Konig L, Maksimovic S, Grass S, Adeberg S, Scherer M, Unterberg A, Bendszus M, et al. CEST imaging of the APT and ssMT predict the overall survival of patients with glioma at the first follow-up after completion of radiotherapy at 3T. Radiother Oncol. 2023;184:109694.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhuo Z, Qu L, Zhang P, Duan Y, Cheng D, Xu X, Sun T, Ding J, Xie C, Liu X, et al. Prediction of H3K27M-mutant brainstem glioma by amide proton transfer-weighted imaging and its derived radiomics. Eur J Nucl Med Mol Imaging. 2021;48(13):4426\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHou H, Chen W, Diao Y, Wang Y, Zhang L, Wang L, Xu M, Yu J, Song T, Liu Y, et al. 3D Amide Proton Transfer-Weighted Imaging for Grading Glioma and Correlating IDH Mutation Status: Added Value to 3D Pseudocontinuous Arterial Spin Labelling Perfusion. Mol Imaging Biol. 2023;25(2):343\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQiu J, Wu K, Zhu M, Chen CY, Luo Y, Liu Y, Wen J. Using quantitative MRI to study the association of isocitrate dehydrogenase (IDH) status with oxygen metabolism and cellular structure changes in glioma. Eur J Radiol. 2022;155:110502.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFigini M, Riva M, Graham M, Castelli GM, Fernandes B, Grimaldi M, Baselli G, Pessina F, Bello L, Zhang H, et al. Prediction of Isocitrate Dehydrogenase Genotype in Brain Gliomas with MRI: Single-Shell versus Multishell Diffusion Models. Radiology. 2018;289(3):788\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao J, Wang YL, Li XB, Hu MS, Li ZH, Song YK, Wang JY, Tian YS, Liu DW, Yan X, et al. Comparative analysis of the diffusion kurtosis imaging and diffusion tensor imaging in grading gliomas, predicting tumour cell proliferation and IDH-1 gene mutation status. J Neurooncol. 2019;141(1):195\u0026ndash;203.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu X, Zhang P, Zhuo Z, Duan Y, Qu L, Cheng D, Sun T, Ding J, Xie C, Liu X et al. Prediction of H3K27M Alteration Status in Brainstem Glioma Using Multi-Shell Diffusion MRI Metrics. J Magn Reson Imaging 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKathrani N, Chauhan RS, Kotwal A, Kulanthaivelu K, Bhat MD, Saini J, Prasad C, Chakrabarti D, Santosh V, Uppar AM, et al. Diffusion and perfusion imaging biomarkers of H3 K27M mutation status in diffuse midline gliomas. Neuroradiology. 2022;64(8):1519\u0026ndash;28.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCalmon R, Dangouloff-Ros V, Varlet P, Deroulers C, Philippe C, Debily MA, Castel D, Beccaria K, Blauwblomme T, Grevent D, et al. Radiogenomics of diffuse intrinsic pontine gliomas (DIPGs): correlation of histological and biological characteristics with multimodal MRI features. Eur Radiol. 2021;31(12):8913\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeibetseder A, Leitner J, Mair MJ, Meckel S, Hainfellner JA, Aichholzer M, Widhalm G, Dieckmann K, Weis S, Furtner J, et al. Prognostic factors in adult brainstem glioma: a tertiary care center analysis and review of the literature. J Neurol. 2022;269(3):1574\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang H, Liu K, Ba R, Zhang Z, Zhang Y, Chen Y, Gu W, Shen Z, Shu Q, Fu J, et al. Histological and molecular classifications of pediatric glioma with time-dependent diffusion MRI-based microstructural mapping. Neuro Oncol. 2023;25(6):1146\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTogao O, Keupp J, Hiwatashi A, Yamashita K, Kikuchi K, Yoneyama M, Honda H. Amide proton transfer imaging of brain tumors using a self-corrected 3D fast spin-echo dixon method: Comparison With separate B(0) correction. Magn Reson Med. 2017;77(6):2272\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou J, Heo HY, Knutsson L, van Zijl PCM, Jiang S. APT-weighted MRI: Techniques, current neuro applications, and challenging issues. J Magn Reson Imaging. 2019;50(2):347\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eToma AK, Holl E, Kitchen ND, Watkins LD. Evans' index revisited: the need for an alternative in normal pressure hydrocephalus. Neurosurgery. 2011;68(4):939\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArab A, Wojna-Pelczar A, Khairnar A, Szabo N, Ruda-Kucerova J. Principles of diffusion kurtosis imaging and its role in early diagnosis of neurodegenerative disorders. Brain Res Bull. 2018;139:91\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu I, Jiang L, Samuelsson ER, Marco Salas S, Beck A, Hack OA, Jeong D, Shaw ML, Englinger B, LaBelle J, et al. The landscape of tumor cell states and spatial organization in H3-K27M mutant diffuse midline glioma across age and location. Nat Genet. 2022;54(12):1881\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoo B, Han K, Ahn SS, Choi YS, Chang JH, Kang SG, Kim SH, Zhou J, Lee SK. Amide proton transfer imaging might predict survival and IDH mutation status in high-grade glioma. Eur Radiol. 2019;29(12):6643\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark JE, Kim HS, Park KJ, Kim SJ, Kim JH, Smith SA. Pre- and Posttreatment Glioma: Comparison of Amide Proton Transfer Imaging with MR Spectroscopy for Biomarkers of Tumor Proliferation. Radiology. 2016;278(2):514\u0026ndash;23.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRegnery S, Adeberg S, Dreher C, Oberhollenzer J, Meissner JE, Goerke S, Windschuh J, Deike-Hofmann K, Bickelhaupt S, Zaiss M, et al. Chemical exchange saturation transfer MRI serves as predictor of early progression in glioblastoma patients. Oncotarget. 2018;9(47):28772\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUxa S, Castillo-Binder P, Kohler R, Stangner K, Muller GA, Engeland K. Ki-67 gene expression. Cell Death Differ. 2021;28(12):3357\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng L, Gong J, Yu T, Zou Y, Zhang M, Nie L, Chen X, Yue Q, Liu Y, Mao Q, et al. Diffuse Midline Gliomas With Histone H3 K27M Mutation in Adults and Children: A Retrospective Series of 164 Cases. Am J Surg Pathol. 2022;46(6):863\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang H, Yang K, Ren X, Cui Y, Li M, Lei Y, Lin S. Diffuse midline glioma with H3 K27M mutation: a comparison integrating the clinical, radiological, and molecular features between adult and pediatric patients. Neuro Oncol. 2020;22(5):e1\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaeda S, Ohka F, Okuno Y, Aoki K, Motomura K, Takeuchi K, Kusakari H, Yanagisawa N, Sato S, Yamaguchi J, et al. H3F3A mutant allele specific imbalance in an aggressive subtype of diffuse midline glioma, H3 K27M-mutant. Acta Neuropathol Commun. 2020;8(1):8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePang H, Dang X, Ren Y, Zhuang D, Qiu T, Chen H, Zhang J, Ma N, Li G, Zhang J, et al. 3D-ASL perfusion correlates with VEGF expression and overall survival in glioma patients: Comparison of quantitative perfusion and pathology on accurate spatial location-matched basis. J Magn Reson Imaging. 2019;50(1):209\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmed","sideBox":"Learn more about [BMC Medicine](http://bmcmedicine.biomedcentral.com/)","snPcode":"12916","submissionUrl":"https://submission.nature.com/new-submission/12916/3","title":"BMC Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"H3K27-altered diffuse midline gliomas, amide proton transfer-weighted imaging (APTw), arterial spin labeling (ASL), diffusion kurtosis imaging (DKI), diagnosis, prognosis","lastPublishedDoi":"10.21203/rs.3.rs-5128033/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5128033/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eThis study aimed to assess the contributions of metabolism, perfusion, and diffusion kurtosis imaging (DKI) to the diagnosis and prognostic prediction of \u003cem\u003eH3K27\u003c/em\u003e-altered diffuse midline gliomas (DMGs).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eBetween June 2020 and May 2023, 95 patients (mean age 16.98 years; 61.1% female) with brainstem tumors, including 71 \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs and 24 \u003cem\u003eH3K27 \u003c/em\u003ewide-type\u003cem\u003e \u003c/em\u003ebrainstem tumors (referred as non-DMGs), underwent preoperative conventional and advanced MRI (amide proton transfer-weighted [APTw], arterial spin labeling [ASL], and DKI). Logistic and Cox regressions with leave-one-out cross-validation (LOOCV) were used to evaluate the separate and integrated contributions of advanced MRI to diagnostic and prognostic tasks.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Advanced MRI techniques significantly improved the diagnostic and prognostic performances for \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs. Specifically, APTw demonstrated predictive value for both diagnosis (odds ratio [OR] = 3.81, p = 0.004) and prognosis (Hazard ratio [HR] = 1.89, p = 0.003). ASL-derived relative cerebral blood flow (rCBF) improved diagnostic (OR = 7.74, p = 0.019) and prognostic (HR = 2.87, p = 0.002) performances. DKI-derived mean diffusivity (MD) was significantly associated with the diagnosis of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs (OR = 0.03, p = 0.012). Integration of these metrics revealed that APTw (OR = 3.41, p = 0.014) and MD (OR = 0.05, p = 0.049) were independent diagnostic variables of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs, while APTw (HR = 1.66, p = 0.037) and rCBF (HR = 2.15, p = 0.038) were independent prognostic factors for overall survival (OS).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eMetabolic, perfusion, and diffusion imaging can improve the diagnosis and prognosis of \u003cem\u003eH3K27\u003c/em\u003e-altered DMGs beyond conventional MRI, which may aid clinical decision-making.\u003c/p\u003e","manuscriptTitle":"Metabolic, Perfusion, and Diffusion Imaging Enhance Diagnosis and Prognosis of H3K27-Altered Diffuse Midline Gliomas","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-01 05:47:17","doi":"10.21203/rs.3.rs-5128033/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-10T11:42:06+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-13T19:00:09+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"301026357063900153154859540084960583385","date":"2025-06-13T18:50:48+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-13T13:40:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"131049448044690718333745274541464252185","date":"2025-06-12T00:55:55+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-06-08T11:23:40+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-06T06:49:01+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medicine","date":"2025-06-05T14:33:50+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmed","sideBox":"Learn more about [BMC Medicine](http://bmcmedicine.biomedcentral.com/)","snPcode":"12916","submissionUrl":"https://submission.nature.com/new-submission/12916/3","title":"BMC Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"04bcd777-0192-43ac-952d-b9ba8d5f19f7","owner":[],"postedDate":"September 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-11-24T16:00:00+00:00","versionOfRecord":{"articleIdentity":"rs-5128033","link":"https://doi.org/10.1186/s12916-025-04472-6","journal":{"identity":"bmc-medicine","isVorOnly":false,"title":"BMC Medicine"},"publishedOn":"2025-11-19 15:56:53","publishedOnDateReadable":"November 19th, 2025"},"versionCreatedAt":"2025-09-01 05:47:17","video":"","vorDoi":"10.1186/s12916-025-04472-6","vorDoiUrl":"https://doi.org/10.1186/s12916-025-04472-6","workflowStages":[]},"version":"v1","identity":"rs-5128033","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5128033","identity":"rs-5128033","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}