Minchul Kim

An interpretable multimodal deep learning model integrating MRI with clinical and molecular variables enabled survival prediction with time-dependent model interpretability in glioblastoma, with imaging-derived features providing complementary prognostic information.

The subventricular zone (SVZ) has been involved in glioblastoma (GBM) progression; however, the detection of mutated cells in the SVZ via imaging remains u

To explore how molecular glioblastoma (mol-GBM) differs from histological glioblastoma (hist-GBM) in tumor heterogeneity using multi-parametric physiologic MRI-based tumor habitat analysis. In this multi-institutional retrospective study, imaging data were collected from two tertiary centers: 13 mol-GBMs and 39 hist-GBMs from institution 1 (2007-2024) for habitat definition, model development, and internal validation and nine mol-GBMs were obtained from institution 2 (2020-2024) for external validation. The apparent diffusion coefficient (ADC; cellularity), relative cerebral blood volume (rCBV; vascularity), and volume transfer constant (K^trans; permeability) were binarized into high/low categories, which yielded eight spatial habitat clusters to visualize tumor heterogeneity. Habitat proportions and intratumoral heterogeneity features were compared between groups. Multivariable logistic regression models incorporating habitat-derived features were developed to distinguish mol-GBM from hist-GBM, and their performance was evaluated. Fifty-two patients (hist-GBM, n = 39; mol-GBM, n = 13; mean age, 60.0 ± 11.1 years; 22 male) were evaluated. Habitat analysis revealed that mol-GBM had a significantly lower proportion of the most malignant habitat (cluster 3: low-ADC, high rCBV, high K^trans; 0.80% vs. 7.0%, P < 0.001). The total proportion of high K^trans habitats was markedly lower in the mol-GBM group (5.7% vs. 21.3%, P < 0.001). A habitat-based multivariable model incorporating tumor volume, cluster 3 proportion, high K^trans proportion, and Shannon entropy achieved an area under receiver operating characteristic curve (AUC) of 0.87 (95% confidence interval [CI]: 0.73-0.97) at internal validation using a leave-one-out cross-validation, which was substantially greater, albeit nonsignificant, than the AUC for tumor size alone (0.70, 95% CI: 0.51-0.85, P = 0.057), with an external validation accuracy of 88.9% (8/9). Multiparametric physiologic MRI habitat analysis demonstrated differences in the tumor heterogeneity between mol-GBM and hist-GBM. Tumor permeability (K^trans) was the most discriminating parameter. Incorporating habitat-derived features alongside tumor volume may improve the discrimination of the two subtypes on MRI analysis. Further studies, including more rigorous external validation, are warranted.

Background: Cognitive reserve (CR) refers to the discrepancy between brain pathology and observed cognitive decline. While education is a key indicator of CR, its role as a potential moderator in the relationships between brain morphology and cognitive impairments in Parkinson's disease (PD) remains unclear. This study examined whether education affects the relationship between brain age and cognitive impairments in patients with PD.Methods: Data from 58 patients with PD were analyzed using a secondary dataset from the OpenNeuro database. Participants aged ≥55 years were on stable medications and underwent standardized neuropsychological assessments. Brain age predictions were generated from T1-weighted magnetic resonance imaging (MRI) using the brainageR package, and the brain age difference (BAD) was calculated after correction for regression dilution. The moderation effect of education on the relationship between BAD and cognition was assessed using Hayes' PROCESS macro. The primary outcome was cognitive performance across six domains: attention, executive function, language, learning and memory, visuospatial ability, and global cognition.Results: Among the six domains, a significant moderation effect of education was found only for language ability (β = 0.01, p = 0.013, R² = 0.20). The relationship between BAD and language was steeper at lower education levels. No statistically significant moderation was found in the remaining five domains.Conclusion: Having more years of education is associated with buffering the effects of accelerated brain aging on language ability in PD.

Purpose This study aimed to evaluate the technique of performing percutaneous radiologic gastrostomy (PRG) by directly puncturing the stomach under ultrasound (US) guidance in patients in whom nasogastric (NG) tube placement was not feasible. Methods Between July 2013 and July 2024, 39 US-guided PRG procedures were attempted in 37 consecutive patients without NG tubes. The initial gastric puncture was performed under US guidance using a 21-22-gauge Chiba needle. The needle was then exchanged for a 6 F Neff catheter, which was used to distend the stomach. Subsequently, a gastrostomy tube was inserted using the conventional technique. Patient demographics, technical success, clinical success, and complications were retrospectively evaluated using electronic medical records. Complications occurring within 30 days of the procedure were categorized according to the Society of Interventional Radiology classification. Results Of 39 procedures, 37 (94.9%) were technically successful. In two cases in which the initial US-guided PRG failed, an angiographic catheter was inserted transorally into the stomach to achieve distension, followed by conventional PRG. All 39 procedures (100%) achieved clinical success, and the patients successfully received nutrition via the gastrostomy tube. No immediate complications occurred. One major late complication (intraperitoneal abscess) and three minor late complications (skin infection, n = 2; tube leakage, n = 1) occurred within 30 days of the procedure. These complications were not associated with the US-guided access itself. Conclusion Performing PRG via direct gastric puncture under US guidance is a safe and effective alternative for patients without an NG tube.

The glymphatic system is a glial-based perivascular network that facilitates the clearance of metabolic waste from the brain. Dysfunction of the glymphatic system, along with neurological symptoms such as cognitive deficits and olfactory dysfunction, has been reported in patients with coronavirus disease (COVID-19). However, the link between these neurological symptoms and alterations in the glymphatic system remains unclear. In this study, we aimed to evaluate magnetic resonance imaging (MRI)-based measures of the glymphatic system in patients recovered from COVID-19 with and without neurological symptoms. This study included 89 patients who recovered from respiratory infections, of whom 71 had confirmed COVID-19 (20 experienced anosmia and 41 had cognitive symptoms). Three MRI-based measures were quantified and compared: the dilated perivascular spaces (dPVS), free water (FW) fraction, and diffusion tensor image analysis along the perivascular spaces (DTI-ALPS). A partial correlation network was used to assess the relationships between COVID-19 infection, neurological symptoms, and glymphatic measures. COVID-19 patients with anosmia had increased FW in the left orbitofrontal area compared to those without anosmia (mean difference: 0.01, p = 0.48), while patients with cognitive symptoms showed decreased left-sided DTI-ALPS (mean difference: 0.06, p = 0.40). Neurological symptoms mediate the relationship between COVID-19 and glymphatic system measures. Our findings imply that neurological symptoms accompanied by COVID-19 are linked to distinct alterations in the glymphatic system, suggesting a potential association between neuroinvasion and neuroinflammatory processes related to COVID-19. • Previous literature indicates alterations in the glymphatic system following COVID-19 infection. • We explore the differences between post-COVID patients with and without neurologic symptoms using MRI-based measures of the glymphatic system. • We demonstrate that post-COVID patients with neurologic symptoms exhibit altered MRI-based measures of the glymphatic system. • These findings suggest that glymphatic alterations associated with COVID-19 may involve direct neuroinvasion and neuroinflammatory processes.

Objectives: To investigate individual variations and outliers in cortical thickness among non‐alcoholic fatty liver disease (NAFLD) patients, ranging from hepatic steatosis to fibrosis, using neuroanatomical normative modeling.Materials and Methods: A cross‐sectional study with 2637 health check‐up subjects was conducted. Among NAFLD patients, hepatic steatosis (n = 556) and fibrosis (n = 57) were determined by hepatic steatosis index and fibrosis‐4 score, respectively. Cortical thickness in 148 different brain regions was assessed using T1‐weighted MRI scans. A publicly available neuroanatomical normative model analyzed cortical thickness distributions with data from around 58,000 participants. The hierarchical Bayesian regression was used to estimate cortical thickness deviation for each region, taking age, sex, and sites into account. On the basis of a normal adaptation set, Z‐scores below −1.96 or above +1.96 per region were classified as outliers. The total outlier count (tOC) was then calculated to quantify regional heterogeneity. Mass univariate analysis was conducted to compare steatosis and fibrosis groups, and the spatial patterns of regional heterogeneity were qualitatively analyzed.Results: Patients with hepatic fibrosis had a higher number of positive outlier regions (mean 6.3 ± 10.3) than hepatic steatosis (mean 4.2 ± 6.2, p = 0.02). Mass univariate group difference testing of 148 brain regions revealed patients with hepatic fibrosis had 6 cortical areas thicker than hepatic steatosis. Two groups showed shared regional heterogeneity in the temporal cortex.Conclusion: Distinct brain atrophy patterns were observed in NAFLD patients compared to the normal group, with more frequent temporal cortex outliers in both hepatic steatosis and fibrosis. Hepatic fibrosis showed slightly increased cortical thickness relative to steatosis.