MRI Scans Research Articles (Page 2)

The hippocampus, a brain region critical for memory, undergoes significant age-related changes at both the macroscopic and microstructural levels. This study investigates these changes using high-gradient diffusion MRI (dMRI) data analyzed in an unfolded hippocampal space. We applied the Soma and Neurite Density Imaging (SANDI) model to quantify microstructural alterations in 72 cognitively healthy participants aged 19-85 years, scanned on a 3 T Connectome MRI scanner with a maximum gradient strength of 300 mT/m. By combining SANDI with a super-resolution algorithm and the HippUnfold toolbox, we achieved high spatial fidelity in our analysis. We observed significant age-related reductions in soma fraction and soma radius, particularly in the subiculum and dentate gyrus, alongside increases in extracellular diffusivity and extracellular fraction, indicating a decline in cellular density and structural integrity. These microstructural changes occur alongside macroscopic alterations such as reduced hippocampal volume and cortical thickness, decreased gyrification, and increased curvature in specific subfields. The spatial correlations between microstructural and macroscopic metrics across the unfolded hippocampal space are weak, both in their mean values and in how they change with age. Our findings suggest that SANDI metrics provide sensitive and complementary information to traditional structural measures, offering new insights into the microstructural underpinnings of hippocampal aging. This study highlights the potential of advanced dMRI techniques to detect subtle age-related changes in hippocampal microstructure, which may contribute to our understanding of aging and its impact on memory and cognition.

Breast cancer is the leading cause of cancer-related death in women worldwide1. Here, in the Breast Cancer-Anti-Progestin Prevention Study 1 (BC-APPS1; NCT02408770 ), we assessed whether progesterone receptor antagonism with ulipristal acetate for 12 weeks reduces surrogate markers of breast cancer risk in 24 premenopausal women. We used multilayered OMICs and live-cell approaches as readouts for molecular features alongside clinical imaging and tissue micromechanics correlates. Ulipristal acetate reduced epithelial proliferation (Ki67) and the proportion, proliferation and colony formation capacity of luminal progenitor cells, the putative cell of origin of aggressive breast cancers2. MRI scans showed reduction in fibroglandular volume with treatment, whereas single-cell RNA sequencing, proteomics, histology and atomic force microscopy identified extracellular matrix remodelling with reduced collagen organization and tissue stiffness. Collagen VI was the most significantly downregulated protein after ulipristal acetate treatment, and we uncovered an unanticipated spatial association between collagen VI and SOX9high luminal progenitor cell localization, establishing a link between collagen organization and luminal progenitor activity. Culture of primary human breast epithelial cells in a stiff environment increased luminal progenitor activity, which was antagonized by anti-progestin therapy, strengthening this mechanistic link. This study offers a template for biologically informed early-phase therapeutic cancer prevention trials and demonstrates the potential for premenopausal breast cancer prevention with progesterone receptor antagonists through stromal remodelling and luminal progenitor suppression.

Background: Anterior Cruciate Ligament (ACL) reconstruction is a common orthopaedic procedure, the success of which is ultimately affected by the graft healing process. Quantification of graft healing can be performed non-invasively, using signal-intensity (SI) or signal noise quotient (SNQ) on MRI, however, the variable factors affecting graft healing are still being studied. Objectives: To non-invasively evaluate the normal morphology of the ACL graft on MRI and assess factors affecting graft healing post-arthroscopic ACL reconstruction. Method: A single-centre cross-sectional study was performed using MRI scans for assessment of the ACL graft at 6 months to 2 years post-surgery. Signal noise quotient was correlated with tibial tunnel diameter, femoral tunnel diameter, tibial tunnel location (antero-posterior and medio-lateral), femoral tunnel location (high-low and deep-shallow), graft bending angle (GBA) and notch volume. Results: Twenty-four of 42 patients had normal grafts (mean ± standard deviation post-operative time: 10.15 ± 4.38 months). The SNQ levels were highest at the proximal part of the graft. Graft SNQ correlated positively with tibial (p = 0.020) and femoral (p ≤ 0.001) tunnel diameters, tibial tunnel location in the medio-lateral direction (P ≤ 0.001), femoral tunnel location in the high-low direction (p ≤ 0.001) and patients having complications. Graft SNQ correlated negatively with tibial tunnel location in the antero-posterior (AP) direction (p ≤ 0.001). Univariate analysis revealed a significant correlation between SNQ and tibial and femoral tunnel diameter, tibial tunnel location in both AP and medio-lateral directions, femoral tunnel location in high-low direction and patients having complications. Multivariate analysis showed the tibial tunnel location (medio-lateral) and the femoral tunnel location (high-low) as the significant independent factors. Conclusion: Intraoperative factors, predominantly the positions of the tibial and femoral tunnels, are the major factors affecting graft healing. Contribution: This study provides greater awareness regarding the factors affecting graft healing, helps establish MRI as an effective non-invasive post-operative imaging modality, and helps surgeons in providing a better individualised approach to surgery.

Brain tumors are complex and potentially life-threatening conditions that require accurate and timely diagnosis. This study proposes NeuroBlend-3, an explainable and hybrid artificial intelligence (AI) framework for multi-class brain tumor classification using MRI scans. The framework begins with preprocessing steps, including grayscale conversion, resizing to 224×224 pixels, normalization, denoising, and enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE). To increase data variability, five augmented versions of each image are generated through horizontal flip, 15° rotation, zooming, Gaussian blur, and brightness adjustment. Deep features are then extracted using six models: HRNet, VGG16, VGG19, ResNet50, ResNet101, and CNN-LSTM. These features undergo optimization using Principal Component Analysis (PCA) and Recursive Feature Elimination (RFE) to reduce redundancy and improve performance. The optimized features train machine learning (ML) models, including XGBoost, AdaBoost, Bagging, and a custom Tree Selection and Stacking Ensemble-based Random Forest (TSRF). To ensure interpretability, explainable AI (XAI) techniques such as Grad-CAM, Grad-CAM++, and LIME are applied to highlight the regions influencing classification decisions. The combination of CNN-LSTM, TSRF, and RFE demonstrates superior performance across all metrics through extensive experimentation. This best-performing combination is termed NeuroBlend-3. Neuro reflects the neurological focus, Blend denotes the fusion of deep and traditional learning approaches, and 3 signifies the integration of CNN-LSTM, TSRF, and RFE. NeuroBlend-3 offers a robust and interpretable solution, making it highly suitable for clinical decision-making in brain tumor diagnosis. Received: 21 June 2025 | Revised: 9 September 2025 | Accepted: 22 October 2025 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The data will be made available to the corresponding author upon request. Author Contribution Statement Mohammed Ibrahim Hussain: Conceptualization, Methodology, Software, Formal analysis, Resources, Writing – original draft, Writing – review & editing. Safiul Haque Chowdhury: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Project administration. Muhammad Minoar Hossain: Writing – review & editing, Supervision. Mohammad Mamun: Software, Validation, Formal analysis, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization.

Increasing MRI capacity is of primary importance to both NHS England and individual radiology departments. Consequently, central funding was provided to allow trusts to instal artificial intelligence-enabled image reconstruction (AI-IR) on their MRI scanners, with the stated aim of increasing capacity by two patients scanned per day within a year of installation on a given scanner. This work demonstrates how a two-phase quality improvement (QI) initiative can be followed to increase capacity using AI-IR in a community diagnostic centre (CDC) at Mile End Hospital and an acute trauma centre, the Royal London Hospital, in East London with comprehensive stakeholders’ engagement.The Model for Improvement framework was used. Our pilot study focused on 3 Plan-Do-Study-Act (PDSA) cycles for three anatomies in musculoskeletal (MSK) imaging at our CDC. A second, substantive study at our major trauma centre was followed, which was a 20-month project encompassing all MSK anatomies of interest.In our initial pilot study at the CDC, we were able to reduce booking times by 10 min for Knee, Ankle and Spine protocols. In our wide-ranging MSK programme at our trauma centre, we saved on average of 07:26 min per scan and while an increased throughput was not achieved, an increase in complex patients being scanned, from 7% to 15% was achieved, reducing healthcare inequities.Our two-centre study suggests that engaging with stakeholders in a structured QI programme can significantly reduce scanning times, improve patient experience and allow for longer precare and postcare time. Additionally, significant throughput increase at the CDC for low-risk ambulatory patients suggests efforts to increase capacity using this technology should be focused at such centres and other scanners focused on ambulatory outpatients, while for scanners focused on inpatients, paediatrics and A&E at trauma centres, the time saved can be used to increase the capacity for complex patients, reducing waiting times for these patients.

The early and accurate identification of brain abnormalities plays a vital role in improving patient outcomes and treatment planning [1], [2]. This project focuses on developing an intelligent medical image analysis system capable of detecting and classifying brain tumors automatically from MRI data [3], [4]. The proposed approach utilizes advanced three-dimensional convolutional neural network (3D-CNN) architectures that effectively capture spatial and contextual information from volumetric MRI images [5]–[7]. The system undergoes preprocessing steps such as skull stripping, normalization, and data augmentation to enhance input quality and model robustness [8], [9]. Through deep feature extraction and layer-wise learning, the model distinguishes between tumor and non-tumor regions with high precision [10], [11]. Experimental results demonstrate that the proposed deep learning framework outperforms conventional 2D models by leveraging 3D spatial relationships within the MRI scans [12]–[15]. This automated solution significantly reduces diagnostic time, assists radiologists in clinical decision-making, and contributes to improved brain healthcare through intelligent image-based diagnosis [16]–[19]. Furthermore, the integration of explainable AI techniques provides interpretability and transparency, which are crucial for clinical trust and real-world applicability [20], [25].

Introduction: Identifying asymptomatic individuals and treating them based on underlying risk is a key challenge in preventing coronary artery disease (CAD). Genetic risk scores and plaque quantification from cardiac imaging have emerged as powerful tools to expand conventional risk stratification. However, these modalities have not been combined in a single predictive model. Objectives: First, to evaluate whether a multimodal AI model integrating imaging, genetic, and lipid-based risk improves prediction of 10-year incident CAD beyond clinical models. Second, to assess whether genetic risk adds predictive value after accounting for imaging. Third, to determine whether non-cardiac imaging modalities contribute independent information. Methods: We analyzed data from over 60,000 UK Biobank participants with ~4,000 CAD events after imaging. Vision models were fine-tuned on cardiac, liver, and pancreas MRI and DXA scans. Imaging embeddings were reduced using principal component analysis and integrated with a multi-ancestry PRS (trained on >2M individuals), metabolic and ECG traits, and baseline variables in a unified Cox proportional hazards model. Model performance was assessed using pseudo R 2 (leave-one-out) and commonality analysis. Results: Imaging embeddings outperformed hand-crafted image-derived phenotypes (AUC: 0.794 vs. 0.666). In joint models, only cardiac long-axis and aortic distensibility MRI contributed substantial independent value; liver, pancreas, and DXA features added minimal predictive power after adjusting for baseline traits. PRS alone explained pseudo R 2 = 0.08, while the full multimodal model reached 0.45, with imaging contributing nearly three times the incremental variance explained by genetics. Genetic and imaging signals were largely orthogonal, though some genetic risk was partially captured by imaging. A hierarchical stratification framework combining clinical, genetic, and imaging data identified a subgroup with a 10-fold increased CAD risk relative to the low-risk baseline and a 5-fold increase compared to individuals with high clinical and genetic risk. Spatial cross-validation confirmed generalizability across imaging centers (AUC: 0.785-0.822 and C-index 0.751-0.763). Conclusions: Genetic risk offers a fixed baseline of inherited susceptibility, but deep learning on non-invasive imaging adds dynamic markers of disease progression. Multimodal modeling offers a practical framework for precision CAD screening at population scale.

Kidney, liver and cyst volumes are important for diagnosis, classification and management of autosomal dominant polycystic kidney disease (ADPKD) but challenging to measure accurately and reproducibly. Here, we develop a web-based deep learning platform to automatically and robustly measure kidneys, liver and cyst volumes in ADPKD. MRI and CT scans from ADPKD patients (n=611) and participants without ADPKD (n=109) were used to train a 3D hybrid model combining U-Net and transformer elements for segmenting kidneys, liver and cysts. The model is implemented as a web-based calculator at www.traceorg.com, providing segmentation labels, volumes and Mayo Clinic Image Classification (MIC). Automatic browser anonymization of DICOM images ensures privacy. Internal validation was conducted on 70 MRIs for kidney and liver segmentations, 46 MRIs for cyst segmentations and performance was compared to 5 open access segmentation models (TotalSegmentator, MR Annotator, Kim, Woznicki and Gregory-Kline). External validation was performed on one single-center dataset (n=58), one multicenter dataset (n=73), CRISP2 (n=30) and PKD-RRC (n=115) MRIs with T2-weighted and T1-weighted images. After training on 720 participants (mean age=48±15, eGFR=74±32 ml/min/1.73m2 and htTKV=826±772ml/m), TraceOrg internal validation performance achieved high mean Dice scores of 0.97 (kidneys), 0.97 (liver), 0.93 (kidney cysts) and 0.82 (liver cysts) outperforming existing models for ADPKD. External validation showed strong performance with Dice scores of 0.92-0.94 (kidney), 0.87-0.96 (liver), 0.85 (kidney cysts) and 0.76-0.90 (liver cysts) for the single-center and 0.95 (kidney), 0.81 (kidney cysts) for the multicenter dataset. Compared to CRISP volumes measured by stereology, mean absolute percent difference was 5.3% (kidneys, n=30), 11% (kidney cysts, n=30) and 5.5% (liver, n=22). Compared to PKD-RRC (n=115), mean absolute percent difference in TKV was 4.9%. TraceOrg is a publicly available web-based tool that automatically measures kidney, liver and cyst volumes from abdominal MRI in ADPKD with high accuracy compared to manual segmentations.

Introduction: Inherited cardiomyopathies and channelopathies can be clinically challenging to diagnose and manage. SCN5A mutation-associated Multifocal Ectopic Premature Purkinje-related Complexes (MEPPC) syndrome is less common. Case Description: A 33-year-old man with WPW syndrome status post ablation, hypertension, and paroxysmal atrial fibrillation presented to the emergency room with dizziness, decreased exercise tolerance, and fatigue. Blood pressure was 138/107 mmHg, heart rate 138 bpm, and normal oxygen saturation. Troponin I was negative, and EKG showed new wide complex tachycardia, LBBB, multifocal PACs, and PVCs. The echocardiogram noted severe LVH, diffuse hypokinesis, and a newly reduced LVEF of 15-20%, with LVEDD of 6 cm. Cardiac MRI had no evidence of infiltrative disease, infarct, or fibrosis. Exercise SPECT was negative for ischemia. He was discharged on a beta blocker, ACE inhibitor, and MRA. He had recurrent VT and PVCs despite being on amiodarone. Atrial fibrillation was controlled after PVI ablation. LV function remained <35%, and a dual-chamber ICD was implanted for primary prophylaxis. Treatment/Outcomes: He had incessant PVC and VT, with multiple ICD shocks on follow-up, leading to the ablation of three PVC morphologies localized to the lateral RVOT and the inferolateral RV. No VT was inducible on attempted VT ablation. Repeat cardiac MRI and PET scan were negative for infiltrative disease. Genetic testing revealed a heterozygous SCN5A mutation with a clinical diagnosis of MEPPC. Given his recurrent VT and cardiomyopathy, he was listed as a Status 6 after cardiac transplant evaluation. He was on maximally tolerated doses of all four GDMT pillars and Flecainide. A follow-up echocardiogram after 6 months of treatment showed a recovered LVEF of 55-60%. His transplant listing was removed due to recovery. He has not experienced any more VT episodes. Discussion/Conclusion: MEPPC is a rare, inherited, autosomal dominant SCN5A-related cardiac syndrome. The mutation results in a gain-of-function of the sodium channel, leading to hyperexcitability of the fascicular-Purkinje system. It is often associated with dilated cardiomyopathy and can lead to sudden cardiac death. Genetic counseling and testing are important in diagnosis. Treatment of MEPPC includes the use of Flecainide. LV dysfunction requires optimal GDMT. In this case, complete LV function recovery was achieved with excellent rhythm control on flecainide along with maximally tolerated GDMT.

This paper presents a multi-agent AI system designed to provide accurate diagnostic and personalized treatment recommendations for heart attack, heart failure, cardiac arrhythmia, coronary artery disease, and left ventricular hypertrophy. The system tackles the challenges of integrating various data sources, including electronic health records (EHR), cardiac imaging, genetic information, and electrocardiogram (ECG) data, within a unified multi-agent framework for personalized care related to these conditions. A collaborative network of specialized AI agents, such as the EHR Agent, Cardiac Imaging Agent, Genetic Analysis Agent, and ECG Analysis Agent, work in concert to process and analyze this multi data, identifying potential cardiac conditions and risk factors associated with the above-mentioned target indicators. Research Questions/Hypothesis: This study investigates whether a multi-agent AI system can effectively process patient data, including symptoms, genetic information, and test results, to generate potential conditions and diagnoses. We hypothesize that this integrated approach can potentially improve the speed of assessment for accurate and timely diagnosis, provide relevant diagnostic information and personalized treatment recommendation. Methods/Approach: The multi-agent system comprises several specialized agents responsible for tasks such as symptom analysis, diagnosis, and treatment planning. The system is targeted at processing patient data, including symptom descriptions and test results from labs (biomarkers), ECG, echo, MRI and CT scans, along with genetic variants. The symptom analysis agent identifies potential cardiovascular conditions based on input symptoms. The diagnostic agent then integrates information from potential conditions, patient history, and test results to generate a diagnosis. Results/Data: Analysis of simulated data demonstrates that the symptom analysis agent consistently identifies expected potential conditions with high level of speed and accuracy. Recording 1-2 seconds of diagnosis time with precision level of 98% based on simulated data and programmed logic. We’re only reporting metrics based on the internal consistency of the agent's logic and simulated outcomes. Conclusion(s): The developed multi-agent system demonstrates a functional approach to integrating diverse simulated patient data for cardiovascular assessment and potential diagnosis.

Background: Electroanatomic mapping (EAM) guidance for endomyocardial biopsies (EMB) has been suggested to be feasible and safe, but the diagnostic yield remains unclear for cardiomyopathies. Objective: We aimed to evaluate the diagnostic efficacy of EAM and intracardiac echocardiography (ICE)-guided EMBs. Methods: We retrospectively reviewed patients who underwent EMB from August 2018 to July 2024. EMB was guided by EAM using CARTO system (Biosense Webster, Irvine, CA), and ICE. After accessing the right femoral vein, samples (3-6 per suspected site) were collected using a disposable bioptome and steerable sheath. For left ventricular (LV) biopsies, we used a transseptal approach, and for atrial biopsies, we targeted the atrial septum. In cases of abnormal EAM, multiple samples were taken from the identified areas. When EAM was normal, biopsy targeting was guided by adjunctive imaging. EMB was considered positive if pathology demonstrated findings that directly corroborated the diagnosis. Results: Of 87 patients who underwent EMB, the median age was 61 years, and 33% were female. EMB sites included the right ventricle (RV) and LV (15/87), RV only (27/87), LV only (38/87), right atrium (RA) and LV (3/87), and RA only (4/87). Pre-procedural imaging was common: cardiac MRI (80%), cardiac PET (65.6%), and/or pyrophosphate scan (8%). Mean LV ejection fraction was 44%, and mean scar burden was 11% on MRI. The overall diagnostic yield was 18%, encompassing a wide spectrum of pathologies (Figure 1A). Positive biopsy results were significantly associated with pre-procedural suspicion of amyloidosis (Odds Ratio {OR} 6.5, 95% CI 1.2-35.5), myocarditis (OR 6.5, 95% CI 1.2-35.5), or cardiac masses (OR 3.9, 95% CI 1.1-13.9), and sampling from both RA and LV (Figure 1B). EAM and ICE during EMB (Figure 1C) were used in 85% and 99% of cases, respectively. No procedural complications were observed. Conclusions: In our cohort, EAM-guided EMB is a safe diagnostic tool with the best yield for pre-procedural suspicion of amyloidosis, myocarditis, or cardiac masses. Future studies investigating the role of potential tools to optimize biopsies for undifferentiated cardiomyopathies and cardiac sarcoidosis could significantly improve the potential value of EAM-guided EMB.

Purpose: Evidence suggests that mast cells (MCs) become overstimulated during reperfusion after a period of prolonged myocardial ischemia. However, the role of MC degranulation on post-reperfusion hemorrhage formation/resolution is unknown. Notably, recent in vitro studies have shown that, as opposed to resting/unstimulated MCs, activated MCs act as "scavengers" by actively engulfing and clearing damaged/oxidized erythrocytes (oxRBC). To date, however, the interaction of MCs with stagnant blood/oxydized erythrocytes in reperfused infarcted myocardium (MI) remains grossly unexplored. In the present study, we investigated the effects of MC stabilization on post-MI hemorrhage resorption/clearance in a clinically relevant porcine model using longitudinal MRI. Methods: Female farm pigs (n=14; 30-35kg) underwent 90-minute occlusion of the left anterior descending artery followed by reperfusion. At Day 5, MI and intramyocardial hemorrhage (IMH) were confirmed using a clinical 3T MRI scanner (Figure 1a). Pigs with comparable MI, IMH and microvascular obstruction (MVO) sizes were randomized into untreated MI (MI, n=7) and treated (LORA, n=7) groups, with the latter given daily oral loratadine (10mg) until termination. The animals were followed longitudinally through Weeks 4 and 8. Cardiac function was assessed using cine MRI sequences (short-axis, horizontal, and vertical long-axis). MI and MVO were quantified using LGE, while IMH was assessed using T2*-weighted imaging. Results: As seen in Figure 1b, no significant differences were observed in infarct ( p =0.23), MVO ( p =0.21) or IMH ( p =0.11) sizes between the untreated MI and LORA groups at Day 5. Hemorrhage resorption from Day 5 to Week 4 and Week 8 was significantly greater in LORA (Week 4: -89.37%; Week 8: -96.53%) than in untreated MI (Week 4: -73.67%, p <0.05; Week 8: -88.66%, p =0.02). Notably, at Week 8, while both groups demonstrated similar ( p =0.48) reductions in MI scar size, the extent of ventricle volume increase was significantly lower in LORA (ESV: +22.23%; EDV: +25.06%) compared to untreated MI (ESV: +64.55%, p <0.01; EDV: +64.11%, p =0.02) group. Conclusions: MC stabilization accelerates hemorrhage resolution and attenuates adverse remodeling in hemorrhagic MI. However, whether overstimulated MC during myocardial ischemia/reperfusion directly exhibit reduced oxRBC scavenging potential, or it is the MC degranulation that indirectly inhibits hemorrhage resolution, remains to be determined.

Description of Case: A 37-year-old woman with hypertension, polycystic ovarian syndrome, and a family history of eosinophilic esophagitis presented 8 weeks postpartum with progressive dyspnea and hypoxia. Laboratory testing showed mild eosinophilia and elevated BNP with a normal troponin. Transthoracic echocardiography revealed preserved ejection fraction and apical thickening. Cardiac MRI (1.5T) demonstrated reduced LVEF (43%), mid-to-apical akinesis, and a large laminar apical thrombus (18 × 51 mm). Native T1 and extracellular volume were elevated. First-pass perfusion showed an apical perfusion defect surrounded by hyperenhancing myocardium, consistent with the “double V sign.” Late gadolinium enhancement demonstrated mid-wall fibrosis in the basal anteroseptum. Endomyocardial biopsy revealed only mild myocyte hypertrophy without eosinophilic infiltration. Discussion: Eosinophilic myocarditis is an uncommon inflammatory cardiomyopathy that often mimics other etiologies of heart failure and may evade histologic confirmation due to patchy myocardial involvement. In this case, the clinical context and distinct cardiac MRI findings supported the diagnosis despite a non-diagnostic biopsy. Given the imaging and laboratory findings, high-dose corticosteroids and anticoagulation were initiated empirically. At three months, repeat imaging showed improved LVEF (57%), resolution of wall motion abnormalities, and thrombus shrinkage to 7 mm. By ten months, MRI demonstrated normalized systolic function (LVEF 60%), near-complete resolution of the thrombus (10 × 6 mm, tethered in the chordae), and no residual late enhancement. T1 and extracellular volume remained mildly elevated. This case illustrates the diagnostic and longitudinal value of cardiac MRI in eosinophilic myocarditis, particularly when biopsy is inconclusive, and highlights the utility of serial imaging in guiding and monitoring therapy.

Background: With the advent of stereotactic radioablation for cardiac arrhythmias, accurate non-contact mapping tools are increasingly important. Magnetocardiography (MCG) is promising but historically limited by supercooled sensors, extensive shielding, and long recording durations. A novel magnetic sensor system developed by TDK Corporation may overcome these prior limitations, but rigorous validation against established methods has not been performed. Specifically, no prior study has directly validated this novel MCG system through a three-way comparison with electrocardiographic imaging (ECGi) and a gold-standard pacing location for atrial arrhythmia localization. Hypothesis: We hypothesized that the novel MCG sensor system would perform comparably to ECGi in accurately localizing atrial activation origins, enabling the creation of reliable atrial activation maps. Methods: Six swine (42.2 ± 5.3 kg) underwent placement of pacing wires in the right atrium to simulate focal atrial arrhythmias. Anatomical MRI scans precisely defined the gold-standard pacing lead position (Fig. A) and ECG electrode locations via fiducials. Atrium anatomy was segmented from MRI images and smoothed to create accurate anatomical models. Simultaneous MCG and ECGi signals were recorded during controlled atrial pacing. Latency maps were generated from denoised, beat-averaged signals (Fig. B). Localization accuracy between MCG and ECGi was compared using a paired Wilcoxon signed-rank test. Results: Approximately 1000 P-waves per animal were analyzed. Median absolute localization error was 24.1 mm (IQR 18.6–30.2 mm) for ECGi and 31.0 mm (IQR 23.2–37.3 mm) for MCG (p=ns; Fig. C). Although localization error was numerically higher for MCG, differences were not statistically significant given the limited sample size. Conclusions: Our preliminary results demonstrate the feasibility of using a novel, solid-state MCG sensor system for non-invasive atrial arrhythmia localization. The difference in localization accuracy between ECGi and MCG was not statistically significant in this initial animal cohort. This first-of-its-kind multimodal validation suggests that novel MCG technology may serve as a viable complementary mapping modality, warranting further validation in larger studies.

Background: Apparent diffusion coefficient (ADC) from diffusion-weighted MRI reflects cytotoxic edema and enables neurological prognostication after cardiac arrest. However, current methods often rely on subjective interpretation or manual region-of-interest analysis, leading to inter-observer variability. Even quantitative tools often require time-consuming manual steps. This study aimed to develop and validate a fully automated whole-brain ADC histogram analysis system that eliminates human bias and enables rapid, objective outcome prediction. Method: This single-center retrospective study included adult out-of-hospital cardiac arrest (OHCA) survivors receiving targeted temperature management. 3T MRI scans acquired 72–96 h after ROSC were included. Patients were randomly divided into derivation (70%) and validation (30%) cohorts. MRI data were processed using JLK-ADC, an AI-driven platform that automatically segments brain parenchyma and generates whole-brain ADC histograms. Neurological outcomes were assessed at 6 months using Cerebral Performance Category (CPC 3–5 defined as poor). Results: A total of 119 comatose OHCA patients were included: derivation cohort (n = 83), validation cohort (n = 36). Voxel-wise histogram analysis revealed that patients with poor neurological outcomes exhibited significantly higher proportions of voxels in low ADC ranges (<=600 x 10^-6 mm2/s), reflecting cytotoxic edema. The 550–600 interval demonstrated highest prognostic performance (AUC 0.792; 95% CI, 0.679–0.891). Cumulative analysis showed that several cutoffs — particularly <=500, <=550, <=600, <=650, and <=700 x 10^-6 mm2/s — were all associated with strong outcome discrimination. The <=600 threshold yielded best overall performance in the derivation cohort at a voxel proportion cutoff of 3.79%, with sensitivity 76.2%, specificity 80.5%, PPV 80.0%, and NPV 76.7%. Internal validation confirmed robust performance: the same <=600 threshold achieved an AUC of 0.840, sensitivity 66.7%, specificity 94.4%, and PPV 92.3%. Pairwise ROC comparisons showed no significant differences in AUCs (p = 0.61), supporting generalizability. Conclusion: This study presents the first fully automated whole-brain ADC histogram analysis for neurological prognostication in cardiac arrest survivors. This approach achieved robust performance across independent validation cohorts, offering clinicians an objective, rapid alternative to subjective manual analysis in post-cardiac arrest care.

Introduction: Quantifying adiposity, a key biomarker of metabolic health, typically requires imaging that involves radiation, high costs, and manual effort. We developed an AI framework to segment external white adipose tissue (EWAT) from routine non-contrast MRI, offering a radiation-free, low-effort alternative. Hypothesis: We hypothesized that combining classical image processing with deep learning would enable accurate, robust EWAT segmentation from routine T1/T2-weighted MRI, without specialized sequences or manual labeling. Methods: In 105 Type 1 diabetes patients, T1/T2-weighted axial abdominal MRI scans at the aortic bifurcation were used to develop three segmentation approaches: Region Growing with automatic seed selection, iterative pixel aggregation and adaptive thresholds; UNet CNN trained on 52 masks from region-growing results, with Dice&Binary Cross-Entropy loss; and, Fine-Tuned UNet, optimized on 48 complex cases using extensive augmentations (flips, crops, brightness shifts, Gaussian noise) to enhance robustness and generalizability. Key challenges like artifacts, low fat volume, and anatomical overlap were addressed via local adjustments and hyperparameter tuning. Three independent clinicians scored segmentation quality (0–3) for anatomical alignment (Accuracy), circumference capture (Completeness), target area segmentation (Coverage), and boundary continuity (Smoothness). Results: Table 1 summarizes the mean clinical evaluation scores across all patients and metrics. The UNet and Fine-Tuned UNet consistently outperformed Region Growing in all four metrics, with mean accuracy of 2.81 and 2.80, respectively, versus 2.16 for Region Growing. Figure 1 shows example segmentations for each method. Figures 2 and 3 visualize model performance, highlighting mean scores in complex cases and the percentage of perfect (3/3) segmentations. The Fine-Tuned UNet had the highest mean accuracy (2.80) in challenging images, while the base UNet had the most perfect scores overall (73.3%). Conclusion: This unsupervised AI framework enables accurate, radiation-free EWAT segmentation from routine MRI. All methods, including deep learning, were trained without manual labeling, using region-growing outputs as pseudo ground truth. Clinical evaluations confirmed that the UNets achieved superior accuracy, completeness, coverage, and smoothness, particularly in complex cases. This scalable, cost-effective approach supports broader validation in cardiometabolic populations.

Opioid use disorder (OUD) is associated with structural brain alterations. Buprenorphine maintenance treatment (BMT)'s impact on brain morphology remains underexplored. We examined the effect of BMT on surface-based morphometry (SBM) metrics- cortical thickness, sulcal depth, gyrification, and fractal dimension, in a longitudinal controlled design. Twenty-five men with OUD and age- and education-matched participants in the control group were recruited. Participants underwent T1-weighted MRI scans immediately after starting BMT and after six months of treatment. SBM metrics were analyzed using the Computational Anatomy Toolbox 12 (CAT12), employing threshold-free cluster enhancement (TFCE) and family-wise error correction. At baseline, individuals with OUD had greater cortical thickness in superior parietal and occipital regions and reduced thickness in the inferior temporal gyrus versus participants in the control group. After six months, significant cortical thickness reductions were observed in the occipital pole, cuneus, and occipito-temporal gyri, and calcarine sulcus in both hemispheres; sulcal depth, gyrification, and fractal dimension remained unchanged. We observed negative correlations between buprenorphine dosage and change in cortical depth in the parahippocampal region (r = -0.53, p = .007) and temporal pole (r = -0.55, p = .005), and positive correlations with fractal dimension in the medial orbitofrontal cortex (r = 0.53, p = .006) and gyrification in the lateral orbital region (r = 0.56, p = .004). BMT is associated with a generalized cortical thinning in sensory regions, while dose-dependent changes are observed in memory, emotional regulation, and cognitive control regions, highlighting neuroadaptive processes in overall treatment and medication-specific effects.

Background: Pediatric pulmonary hypertension can be diagnosed by echocardiography and right heart catheterization, but cardiac MRI-based septal curvature (SC) measurement can also be used as a surrogate of mean pulmonary arterial pressure (mPAP), which is an invasive measurement to follow-up patients. We developed an automated approach to measure SC, demonstrating its superiority over a manual measurement. However, its performance relative to other septal wall measurements and clinical markers are unclear. Hypothesis: Automated SC is better correlated with mPAP, less observer dependent, and better associated with adverse outcomes than interventricular septal angle (IVS) and right ventricular ejection fraction (RVEF). Aims: To compare the automated SC, IVS, and RVEF in terms of observer variability, correlation to mPAP, and correlation with adverse outcomes. Methods: Patients with pulmonary hypertension who had both catheterization and cardiac MRI were retrospectively included. Automated SC and IVS were measured using a mid-slice of short-axis stack imaging for both ventricles using cvi42, a custom MATLAB tool and Fuji PACs (Fig.1). RVEF was collected from the MRI scan report. Adverse outcomes were death, transplant, and/or indication for transplant of heart and/or lung and were collected from electronic health record. Pearson correlation was used for correlation between the metrics and mPAP. A receiver-operating characteristic (ROC) curve was used to investigate the association between the metrics and outcomes. Intraclass correlation coefficient (ICC) was used for interobserver variability analysis. P<0.05 was considered statistically significant. Results: 25 patients (17.0 [12.0 – 18.0] years; 13 with adverse outcomes) were included. Automated SC had a better correlation with mPAP (R=-0.82, p<0.001) than IVS (R=0.66, p<0.001) and RVEF (R=-0.49, p=0.01) (Fig.2). The capability to differentiate adverse outcomes was significant and better for RVEF (area under the curve of 0.82, p=0.007) while it was not significant for automated SC (0.72, p=0.06) and IVS (0.63, p=0.28) (Fig.3). Interobserver analysis found comparable ICCs (0.98, 95%CI, 0.97 – 0.99 for automated SC; 0.97, 95%CI 0.94 – 0.98 for IVS). ICC was not estimated for RVEF due to retrospective nature of the data collection. Conclusion: The automated SC better correlated with mPAP, with comparable observer dependency to IVS but was not able to better differentiate adverse outcomes than RVEF.

Access to prostate MRI remains limited due to resource constraints and the need for expert interpretation. To develop machine learning (ML) models that enable risk-based triage for prostate MRI (ProMT-ML) in the evaluation of prostate cancer. Retrospective and prospective. A total of 11,879 retrospective MRI scans for suspected prostate cancer from a multi-hospital health system, divided into training (N = 9504) and test (N = 2375) sets. A total of 4551 records for prospective validation. 1.5T and 3T/Turbo-spin echo T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE). Prostate Imaging Reporting and Data System (PI-RADS) scores were retrieved from MRI reports. The Boruta algorithm was used to select final input features from candidate features. Two models were developed using supervised ML to estimate the likelihood of an abnormal MRI, defined as PI-RADS ≥ 3: Model A (with prostate volume) and Model B (without prostate volume). Models were compared to PSA. Prostate biopsy pathology was assessed to evaluate potential clinical impact. Area under the receiver operating characteristic curve (AUC) was the primary performance metric. A total of 5580 (46.9%) subjects had a PI-RADS score ≥ 3. After feature selection, Model A included age, PSA, body mass index, and prostate volume, while Model B included age, PSA, body mass index, and systolic blood pressure. Both models A (AUC 0.711) and B (AUC 0.616) significantly outperformed PSA (AUC 0.593). Compared to PSA threshold > 4 ng/mL, Model A demonstrated significantly improved specificity (28.3% vs. 21.9%) and no significant difference in sensitivity (89.0% vs. 86.7%). Among false negatives (Model A: 8.0% (62/776); Model B: 16.8% (130/776)), most (Model A: 87%; Model B: 69%) had benign or clinically insignificant disease on biopsy. On prospective validation, both versions of ProMT-ML significantly outperformed PSA. ProMT-ML provides personalized risk estimates of abnormal prostate MRI and can support triage of this test. Stage 4.

Background: The long-term survival of heart transplantation (HTx) recipients is influenced by a range of cardiovascular, immunological, and procedural factors. Accurately predicting post-HTx outcomes remains a major clinical challenge, especially when relying solely on noninvasive methods. Objective: To test the hypothesis that structural and functional indices derived from multi-parametric cardiac MRI-derived can be used to predict cardiovascular events in HTx recipients. Materials and methods: With the approval of institutional review board (IRB), 170 HTx recipients (106 males, age: 47.8 ± 16 years, Range: 19 – 79 years) were recruited for a comprehensive multi-parametric cardiac MRI scan. MRI images were processed to derive global cardiac function and volumes, and myocardial T2 values and T1 values. Pre- and post-Gadolinium T1 was used to calculate extra-cellular volume (ECV) fraction. Cardiovascular events were defined as a composite of any emergency visit, hospitalization or death due to graft failure or reception, myocardial infarction, HF and other events that cannot rule out a cardiovascular origin of complications. Identification of predictors of adverse outcomes at long-term follow-up was based on a Cox proportional hazards model (CPH). Statistical analysis was performed by using SPSS (version 22.0). Results: MRI images were eligilbe for quantitative analysis. See figure 1. The patients were followed for 6 to 4504 days (Median = 2616 days) after multi-parametric cardiac MRI. In total, 140 cardiovascular events occurred (6 to 3294 days, Median = 627 days). The CPH model fits the data (p < 0.001). After the adjustment of traditional cardiovascular risk factors and demographic data, multiple MRI-derived indices were identified as significant predictors of survival time (time between baseline cardiac MRI and adverse event), including left ventricular (LV) end-diastolic volume (LVEDV) (p < 0.001), LV end-systolic volume (LVESV) (p < 0.001), LV stroke volume (LVSV) (p = 0.005), right ventricular (RV) stroke volume (RVSV) (p < 0.001), RV cardiac output (RVCO) ( p = 0.03), myocardial ECV (p < 0.001) and T2 value (p = 0.008). See figure 2. Conclusions: Multi-parametric indices of cardiac tissue (T2, ECV) and function (LVEDV, LVESV, LVSV, RVSV, RVCO) can independently predict adverse clinical outcomes in HTx recipients at long term follow-up (median > 7 years). MRI may offer new imaging biomarkers for early identification of risks for post-HTx complication.