Neuroradiology最新文献

{"title":"An MRI-based radiomics model for precision subtyping of hereditary spastic paraplegia: discriminating SPG4 from SPG5.","authors":"Zhuoting Huang, Fan Zhang, Longjun Yuan, Yuxian Wan, Xinyu Yan, Xiang Lin, Jianping Hu, Ying Liu","doi":"10.1007/s00234-026-04064-y","DOIUrl":"10.1007/s00234-026-04064-y","url":null,"abstract":"<p><strong>Background: </strong>Differentiating spastic paraplegia type 4 (SPG4) from type 5 (SPG5) is clinically relevant with differing therapeutic strategies, yet genetic testing remains limited necessitate alternative approaches. This study aimed to develop and validate an interpretable radiomics model based on spinal cord MRI for distinguishing SPG4 from SPG5.</p><p><strong>Methods: </strong>In this prospective study, SPG4 (N = 40) and SPG5 (N = 30) were randomly assigned to training (n = 56) and test (n = 14) sets. Radiomic features were extracted from the whole spinal cord (WSC) and cervical spinal cord functional subregions (SCFS) on three-dimensional T1- and T2-weighted images. Predictive models were constructed using four machine learning algorithms and evaluated by receiver operating characteristic curve (ROC) analysis. Model robustness was assessed by permutation testing, and interpretability was provided through SHapley Additive exPlanations (SHAP).</p><p><strong>Results: </strong>Radiomics models based on the WSC consistently outperformed those based on SCFS. The optimal WSC model, constructed with a linear support vector classifier on combined T1- and T2-weighted sequences, achieved an AUC of 1.00 in the test set. SHAP analysis identified five consensus features contributed significantly to discrimination between SPG4 and SPG5. All five features also showed significant differences between the two groups (p < 0.001), with the most influential feature (T1_original_shape_SurfaceVolumeRatio) correlating with disease severity as measured by the Spastic Paraplegia Rating Scale (r = 0.434, p < 0.01).</p><p><strong>Conclusions: </strong>Spinal cord radiomics enables accurate and interpretable differentiation between SPG4 and SPG5, providing a non-invasive adjunct to genetic testing. This approach offers objective imaging biomarkers that may support precision diagnosis, improve patient stratification in hereditary spastic paraplegia.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2963-2972"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148199651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Association between posterior cerebral artery involvement and arterial wall enhancement in moyamoya disease: a cross-sectional study with vessel wall imaging.","authors":"Takeshi Funaki, Hideo Chihara, Tomoki Sasagasako, Masakazu Okawa, Yohei Mineharu, Yasutaka Fushimi, Sachi Okuchi, Kentaro Ueno, John Grinstead, Sinyeob Ahn, Alto Stemmer, Taichi Ikedo, Hiroyuki Ikeda, Yoshiki Arakawa","doi":"10.1007/s00234-026-04037-1","DOIUrl":"10.1007/s00234-026-04037-1","url":null,"abstract":"<p><strong>Purpose: </strong>Early detection and activity assessment of posterior cerebral artery (PCA) involvement in moyamoya disease (MMD) are challenging. We aimed to assess the cross-sectional association between PCA involvement and PCA wall enhancement on vessel wall MRI (DANTE T1-SPACE).</p><p><strong>Methods: </strong>We analyzed consecutive patients with MMD who underwent both non-contrast-enhanced and contrast-enhanced DANTE T1-SPACE. Positive wall enhancement was defined as location-appropriate arterial wall enhancement with an enhancement ratio ≥ 1.5. PCA involvement was graded 0-2 on MRA obtained during the same period.</p><p><strong>Results: </strong>Of 368 hemispheres from 184 patients, 56 (15.2%) showed PCA involvement and 17 (4.6%) showed PCA wall enhancement. PCA involvement was present in 94.1% of enhancement-positive hemispheres versus 11.4% of enhancement-negative hemispheres (p < 0.001). The difference was consistent across age, Suzuki stage, and RNF213 genotype strata. Enhancement most often involved the P2A segment. Topographical correspondence between enhancement and the stenoocclusive lesion was excellent (overall percent agreement, 94.1%). PCA wall enhancement was observed in 0.3%, 61.5%, and 18.6% of hemispheres with grade 0, 1, and 2 PCA involvement, respectively. Positive enhancement was more frequent in grade 1 PCA involvement than in grade 2 (adjusted p = 0.02).</p><p><strong>Conclusions: </strong>Statistical and topographical findings suggest a strong cross-sectional association between PCA involvement and PCA wall enhancement. The enhancement may be associated with disease activity and may facilitate detection of early-stage PCA involvement with high specificity; these hypotheses require confirmation in longitudinal cohorts.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2613-2621"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147948727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hyperdense middle cerebral artery sign is not associated with functional outcome following mechanical thrombectomy.","authors":"Tomislav Herega, Ivan Perić, David Ozretić, Marko Radoš","doi":"10.1007/s00234-026-04129-y","DOIUrl":"10.1007/s00234-026-04129-y","url":null,"abstract":"<p><strong>Purpose: </strong>Arterial hyperdensity on non-contrast CT is a well-recognized indicator of acute thromboembolism within the intracranial arteries. However, its association with functional outcome following mechanical thrombectomy remains inconclusive. This study aimed to evaluate the association between the hyperdense middle cerebral artery sign (HMCAS) and functional outcome in a homogeneous cohort of patients treated with mechanical thrombectomy for occlusion of the M1 segment of the middle cerebral artery (MCA).  METHODS: We conducted a retrospective analysis of a tertiary stroke center prospectively maintained database to identify patients who underwent mechanical thrombectomy for M1 segment occlusion between January 1, 2021, and December 31, 2023. Demographic data, comorbidities, imaging scores, procedural parameters, and clinical parameters were collected.  RESULTS: A total of 220 patients were included in the study. The prevalence of HMCAS did not differ significantly between patients with favorable and unfavorable functional outcomes. Patients with favorable outcomes were significantly younger, had a lower prevalence of arterial hypertension, lower NIHSS scores, better collateral status, and a shorter median symptom onset-to-recanalization time. In multivariable binary logistic regression analysis, HMCAS was not associated with functional outcome (OR 1.41, 95% CI 0.66-3.05, p = 0.377). Lower NIHSS scores and well-developed leptomeningeal collaterals were the strongest predictors of a favorable functional outcome.  CONCLUSION: HMCAS is not associated with functional outcome in patients undergoing mechanical thrombectomy for acute ischemic stroke caused by M1 segment occlusion. Other clinical and collateral factors are the principal determinants of functional outcome.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2567-2576"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148664240","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multimodal MRI assessment of cerebral small vessel disease-related brain injury in the context of unilateral severe carotid artery stenosis.","authors":"Jiawen Gao, Min Wang, Yingying Qiao, Haotian Zhao, Jinyuan Weng, Ning Zheng, Shuo Shao","doi":"10.1007/s00234-026-04111-8","DOIUrl":"10.1007/s00234-026-04111-8","url":null,"abstract":"<p><strong>Purpose: </strong>To investigate the transhemispheric effects of unilateral severe carotid artery stenosis (CAS) on brain structure and hemodynamics, and their association with cerebral small vessel disease (CSVD) markers.</p><p><strong>Methods: </strong>Fifty-one patients with unilateral severe CAS and 48 matched healthy controls underwent multimodal MRI. Quantified metrics included cortical thickness, white matter hyperintensity (WMH) volume, free water fraction in white matter (FW-WM), peak width of skeletonized mean diffusivity (PSMD), and lateralization index for cerebral blood flow (LI-CBF). Group comparisons, interhemispheric comparisons, partial correlation analyses, and multivariable linear regression models were performed.</p><p><strong>Results: </strong>Compared with controls, CAS patients exhibited reduced cortical thickness (2.32 ± 0.08 vs. 2.36 ± 0.08 mm, P = 0.015), elevated PSMD (median 2.56 × 10⁻⁴ vs. 2.35 × 10⁻⁴, P = 0.006), increased WMH volume (median 4.65 vs. 2.37 cm³, P < 0.001), higher FW-WM (0.21 ± 0.03 vs. 0.17 ± 0.02, P < 0.001), and pronounced CBF lateralization (P < 0.001). Ipsilateral FW-WM was significantly higher than contralateral FW-WM (P < 0.001). Both FW-WM (β'=0.347, P = 0.019) and PSMD (β'=0.426, P = 0.002) were independently associated with global WMH volume. Age and anterior cerebral artery LI-CBF independently predicted FW-WM and PSMD. Ipsilateral WMH volume (β'=-0.264, P = 0.044) and age (β'=-0.371, P = 0.005) were independently associated with MoCA scores.</p><p><strong>Conclusion: </strong>Unilateral severe CAS induces structural and hemodynamic alterations overlapping with CSVD features. Hemodynamic disturbance is associated with diffuse white matter injury, which may contribute to WMH formation. Multimodal MRI metrics provide imaging evidence linking macrovascular stenosis to CSVD.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2661-2671"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148423032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning versus conventional grading systems for prognostication in aneurysmal subarachnoid hemorrhage: a systematic review and meta-analysis.","authors":"Donald Ogolo, Obioma Akwada, Enyereibe Ajare, Brenda Opara, Francis Campbell, Okwunodulu Okwuoma, Wilfred Mezue, Samuel Ohaegbulam","doi":"10.1007/s00234-026-04115-4","DOIUrl":"10.1007/s00234-026-04115-4","url":null,"abstract":"<p><strong>Background: </strong>Accurate prognostication after aneurysmal subarachnoid hemorrhage (aSAH) remains challenging. Conventional clinical and radiological grading systems, including the World Federation of Neurosurgical Societies (WFNS), Hunt-Hess, and Fisher scales, are widely used but have limited discriminative capacity. This study aimed to systematically compare machine learning (ML)-based prognostic models with conventional grading systems for predicting functional outcomes and mortality after aSAH, and to evaluate factors influencing ML performance.</p><p><strong>Methods: </strong>A systematic review and meta-analysis were conducted according to PRISMA 2020 guidelines. PubMed, Embase, Scopus, Web of Science, and the Cochrane Library were searched for studies published between 2010 and 2025. Eligible studies evaluated ML-based models for outcome prediction in adult aSAH patients and reported performance of conventional grading systems. Prognostic discrimination was pooled using random-effects meta-analysis of the area under the receiver operating characteristic curve (AUC), with predefined subgroup analyses.</p><p><strong>Results: </strong>Fourteen studies including 6,247 patients were analyzed. ML models demonstrated good to excellent discrimination, with AUCs ranging from 0.81 to 0.97. The pooled ML AUC for predicting unfavourable neurological outcome was 0.86 (95% CI 0.83-0.89; p < 0.0001), with substantial heterogeneity (I² = 96.1%). ML models outperformed conventional grading systems in most studies and showed comparable performance in the remainder. Subgroup analyses confirmed statistically significant prognostic accuracy across clinical-only, imaging-based, and multimodal ML models.</p><p><strong>Conclusion: </strong>Machine learning-based prognostic models demonstrate statistically significant and clinically meaningful performance for outcome prediction after aSAH, exceeding conventional grading systems and supporting their role as complementary risk stratification tools.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2755-2767"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148579273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reduced choroid plexus volume aggravates the glymphatic system burden in patients with patent foramen ovale-related migraine with aura.","authors":"Liqiang Sun, Kaige Cui, Jing Hu, Yu Zhai, Jiping Yang","doi":"10.1007/s00234-026-04094-6","DOIUrl":"10.1007/s00234-026-04094-6","url":null,"abstract":"<p><strong>Objective: </strong>The choroid plexus (ChP) produces cerebrospinal fluid (CSF) and serves as a crucial component of the glymphatic system. This study investigated the characteristics of choroid plexus volume (CPV) changes in patients with patent foramen ovale-related migraine with aura (PFO-M), analyzed its correlation with migraine severity (including both visual analogue scale [VAS] and monthly migraine frequency [MMF]), and elucidated the mechanism of chronic long-term injury on ChP as well as the therapeutic significance of PFO closure.</p><p><strong>Methods: </strong>A prospective study was conducted to enroll 76 patients with PFO-related migraine with aura (PFO group) and 50 healthy controls (HC group) who were treated in our hospital from September 2023 to September 2025. 3.0T magnetic resonance imaging (MRI) was used to measure the lateral ventricular CPV of all subjects, and both VAS and MMF were adopted to evaluate the migraine severity of patients. Intergroup differences in CPV were calculated, and the correlations between CPV and migraine severity indicators (VAS and MMF) in the PFO-M group was analyzed. 26 patients underwent follow-up MRI at 6 months postoperatively to measure CPV and reassess the VAS and MMF, and pre- and post-closure changes in these metrics were compared.</p><p><strong>Results: </strong>(1) The CPV in the PFO-M group was significantly smaller than that in the HC group (1709.76 ± 468.23 mm³ vs. 2443.68 ± 797.48 mm³, t=-5.875, P < 0.001). (2) A significant negative correlation was observed between CPV and VAS (r=-0.788, P < 0.001) and MMF (r=-0.769, P < 0.001) in the PFO-M group. (3) Migraine severity in the PFO-M group at 6 months postoperatively was significantly lower compared with pre-closure levels (VAS: 1.04 ± 1.54 vs. 6.12 ± 1.84, t = 10.718, P < 0.001; MMF: 0.92 ± 1.32 vs. 5.08 ± 1.76, t = 10.316, P < 0.001). (4) There was no statistically significant difference in CPV of the PFO-M group between pre- and 6 months post-closure (1658.23 ± 421.99 mm³ vs. 1666.54 ± 486.86 mm³, t = 0.225, P = 0.823).</p><p><strong>Conclusion: </strong>Patients with PFO-related migraine with aura show reduced choroid plexus volume correlated with migraine severity. PFO closure improves migraine symptoms, without significant changes in choroid plexus volume at 6 months, suggesting longer follow-up is needed.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"3041-3048"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148345774","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Position-dependent white matter microstructural differences in amateur soccer players: a multi-shell diffusion MRI study.","authors":"Demet Küçük, Hamza Küçük, Barış Genç, Kerim Aslan, Hüseyin Akan","doi":"10.1007/s00234-026-04087-5","DOIUrl":"10.1007/s00234-026-04087-5","url":null,"abstract":"<p><strong>Purpose: </strong>To exploratorily evaluate whether long-term exposure to repetitive subconcussive head impacts is associated with position-related patterns of white matter microstructural variation in amateur soccer players using multi-shell diffusion MRI.</p><p><strong>Methods: </strong>This prospective observational study included 43 male amateur soccer players (defenders, n = 17; midfielders, n = 14; goalkeepers, n = 12), aged 18-27 years, each with ≥ 10 years of playing experience. Player position was defined as the primary registered position reported by the participant's club and confirmed by self-report as the predominant playing role over the preceding decade. Participants reporting frequent position switching or hybrid tactical roles were excluded to reduce positional overlap. Diffusion tensor imaging (DTI) and neurite orientation dispersion and density imaging (NODDI) metrics were analyzed using tract-based spatial statistics. Given the modest sample size and the use of player position as an indirect proxy for repetitive head-impact exposure, the analytical framework was defined a priori as exploratory and hypothesis-generating.</p><p><strong>Results: </strong>Defenders demonstrated more widespread white matter microstructural differences compared with midfielders, characterized by reduced fractional anisotropy and neurite density index and increased radial and mean diffusivity across multiple association, projection, and commissural tracts (P < 0.05, TFCE-corrected). Comparisons involving goalkeepers revealed fewer and more spatially restricted alterations. No cerebral microhemorrhages were detected on susceptibility-weighted imaging. Neurocognitive performance did not differ significantly among groups.</p><p><strong>Conclusions: </strong>Within the constraints of an exploratory design and operational position definitions, multi-shell diffusion MRI identified position-associated patterns of white matter microstructural variation in amateur soccer players, with defenders showing more extensive alterations. These findings should be interpreted cautiously and primarily serve to inform future studies employing larger samples, objective exposure metrics, and longitudinal designs.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2783-2792"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148352764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enlarged choroid plexus in chronic insomnia patients and its association with DTI-ALPS index alterations and cognitive decline.","authors":"Junbin Liu, Qianglong Lin, Jingwen Li, Guang Xu, Mengchen Liu, Tianjun Zhang, Xiaofen Ma","doi":"10.1007/s00234-026-04065-x","DOIUrl":"10.1007/s00234-026-04065-x","url":null,"abstract":"<p><strong>Background: </strong>The phenomenon of decreased DTI-ALPS index has been reported in patients with chronic insomnia (CI) and has been associated with cognitive decline. Choroid plexus (ChP) volume is closely linked to glymphatic system function. However, the ChP volume in patients with CI and its relationship with both DTI-ALPS index and cognitive function remain unclear.</p><p><strong>Methods: </strong>The study included 107 patients diagnosed with CI and 68 healthy controls. All participants underwent brain MRI scans, sleep assessments, and a comprehensive assessment of cognitive function. The ChP volume was automatically segmented from 3D-T1WI and presented as a ratio to total intracranial volume. The DTI-ALPS index was calculated based on DTI and used to evaluate water diffusivity in specific white matter regions.</p><p><strong>Results: </strong>Our findings demonstrated significant enlargement of ChP volume and reduction in the DTI-ALPS index among CI patients with declined cognition, compared to patients with normal cognition and healthy controls. The results of partial correlation analysis showed that ChP volume in CI patients with declined cognition was negatively correlated with the Montreal Cognitive Assessment (MoCA) (r = -0.350, P = 0.009) and the Digit Span Test (DST) (r = -0.273, P = 0.044). ChP volume was negatively correlated with DTI-ALPS index (r = -0.342, P = 0.011). Furthermore, DTI-ALPS index was positively correlated with MoCA (r = 0.295, P = 0.029) and DST (r = 0.269, P = 0.047). The results of receiver operating characteristic analysis showed that ChP volume is inadequate as a highly accurate (area under the curve: 0.65, 95% CI: 0.54-0.76, P = 0.009) imaging marker for distinguishing cognitive decline in patients with CI.</p><p><strong>Conclusions: </strong>Our study indicated an enlargement of ChP in CI patients with declined cognition. Enlarged ChP in CI patients with declined cognition was associated with lower DTI-ALPS index and cognitive decline.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2985-2995"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148318427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated deep learning-based segmentation and volumetric analysis of meningiomas.","authors":"Daniel de Wilde, Olivier Zanier, Aron Alakmeh, Stefanos Voglis, Ulf C Schneider, Luca Regli, Carlo Serra, Victor E Staartjes, Gustav Burström, Raffaele Da Mutten","doi":"10.1007/s00234-026-04152-z","DOIUrl":"10.1007/s00234-026-04152-z","url":null,"abstract":"<p><strong>Introduction: </strong>Meningiomas are the most common primary intracranial tumors and are frequently monitored over extended periods. Volumetric assessment typically requires manual segmentation, which is time-consuming and associated with interrater variability. This study aimed to develop and validate a deep learning-based model for the automated segmentation of meningiomas and associated peritumoral edema on preoperative magnetic resonance imaging (MRI).</p><p><strong>Methods: </strong>We trained a standard nnU-Net deep learning model on contrast-enhanced T1-weighted and FLAIR MRI scans from 100 patients treated at the University Hospital of Zurich. The model was then externally validated on 88 cases from the meningioma SEG-Class dataset from the Cancer Imaging Archive. Segmentation performance was assessed using the Dice similarity coefficient, Jaccard index, and 95th percentile Hausdorff distance.</p><p><strong>Results: </strong>The model achieved mean Dice scores of 0.87 ± 0.23 for meningioma segmentation and 0.63 ± 0.38 for peritumoral edema in internal cross-validation. On the external validation set, the model achieved scores of 0.86 ± 0.17 for meningioma segmentation and 0.31 ± 0.35 for edema.</p><p><strong>Conclusion: </strong>The deep learning model demonstrated high accuracy in segmenting meningiomas and modest performance for peritumoral edema. These results support the potential utility of automated segmentation tools in clinical workflows. Future work should focus on validating model performance across larger multi-center datasets.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2917-2926"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13680523/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148796091","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CT-based deep learning radiogenomics for predicting key glioma genotypes (IDH, ATRX, EGFR, TP53).","authors":"Zohal Alnour Ahmed Emam, Emel Ada, Burçin Pehlivanoğlu, Berrin Çavuşoğlu, Kadir Akgüngör, M Alper Selver","doi":"10.1007/s00234-026-04043-3","DOIUrl":"10.1007/s00234-026-04043-3","url":null,"abstract":"<p><strong>Purpose: </strong>Molecular subtyping guides diagnosis and targeted therapy for gliomas. Although MRI-the current imaging standard-can be time-consuming and is sometimes contraindicated, computed tomography (CT) is faster, more widely available, and often preferable in emergency and resource-limited settings. We evaluated whether CT-based radiogenomic signatures combined with machine learning could accurately predict clinically relevant glioma molecular markers.</p><p><strong>Methods: </strong>In this retrospective study, we characterised non-contrast CT (NCCT) scans from 197 adults with histologically confirmed gliomas. Models were developed to predict mutations in four molecular markers: ATRX (n=81), EGFR (n=17), TP53 (n=71), and IDH (n=183). We extracted 208 quantitative radiomic features and added basic demographic variables. Feature selection used LASSO-RFE and Gradient Boosting-RFE with cross-validation. Six classical machine-learning classifiers and deep-learning approaches - including custom fully connected neural networks (FCNN) and TabNet - were trained and compared using ROC-AUC as the primary performance metric.</p><p><strong>Results: </strong>Deep-learning methods outperformed conventional classifiers for all targets. TabNet achieved ROC-AUCs of 0.900 (95% CI: 0.717-0.989; ATRX), 0.955 (95% CI: 0.661-0.978; TP53), and 0.917 (95% CI: 0.858-0.975; EGFR). A custom FCNN obtained a ROC-AUC of 0.971 (95% CI: 0.876-0.995) for IDH. Cross-validation coefficients of variation were 2.0% for ATRX and 3.0% for TP53 and EGFR, and 16.0% for IDH. Deep-learning approaches yielded statistically significant improvements over conventional methods (p-values ranging from <0.05 to <0.001).</p><p><strong>Conclusion: </strong>NCCT-based analytical methods were able to predict clinically relevant genetic mutations in gliomas and demonstrated performance comparable to established techniques. These findings suggest that CT may serve as a practical option for molecular profiling in urgent or resource-limited settings. Nonetheless, external validation is necessary prior to clinical translation. The EGFR findings, arising from a small EGFR-tested subgroup (n = 17), remain preliminary; the observed performance may not reflect generalisable accuracy, and independent validation in larger cohorts is required.</p>","PeriodicalId":19422,"journal":{"name":"Neuroradiology","volume":" ","pages":"2859-2875"},"PeriodicalIF":2.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13680328/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148157100","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}