Results
- Showing results for:
- Reset all filters
Search results
-
Conference paperFang Y, Zhou H, Lee KH, et al., 2027, , Pages: 645-655, ISSN: 0302-9743
Generating precise diagnostic reports from high-resolution computed tomography is critical for clinical workflows, yet remains challenging due to the high pathological diversity and spatial sparsity within 3D volumes. While video language models (VideoLMs) have demonstrated remarkable spatio-temporal reasoning in general domains, their adaptability to domain-specific, high-volume medical interpretation remains underexplored. In this work, we present AbSteering, an abnormality-centric framework that adapts general-domain VideoLMs to precise report generation. Specifically, AbSteering introduces: (i) an abnormality-centric Chain-of-Thought scheme that encourages finding-first abnormality reasoning, and (ii) a Direct Preference Optimization objective that uses clinically confusable abnormalities as hard negatives to enhance fine-grained discrimination. Our results demonstrate that general-purpose VideoLMs exhibit strong transferability to high-volume medical imaging when guided by AbSteering. Notably, VideoLMs outperform state-of-the-art domain-specific CT foundation models, achieving superior detection sensitivity while mitigating hallucinated findings. Our data and model weights are released at https://github.com/ayanglab/CTRG_VideoLMs.
-
Conference paperHu L, Wang Z, Wu Y, et al., 2027, , Pages: 538-548, ISSN: 0302-9743
Cardiac magnetic resonance (CMR) is a clinical reference standard for assessing cardiac morphology and function. Whole-heart assessment relies on sparse multi-plane cine acquisitions from complementary short-axis (SAX) and 2/3/4-chamber long-axis (LAX) views across the cardiac cycle. Although large-scale cohorts such as UK Biobank provide tens of thousands of such studies, most scans remain unlabeled, and learning compact whole-heart representations from the resulting high-dimensional 3D+t data is still challenging under practical memory constraints. To address this, we introduce a geometry-aware self-supervised learning framework for whole-heart representation learning from sparse multi-plane cine CMR. First, we inject patient-space coordinates into spatio-temporal tubelet tokens, which preserves explicit spatial identity and enables shared per-plane encoding. Next, we explicitly model cross-plane structure by aggregating multi-plane features into a latent 3D volume via voxel-query cross-attention. To ensure geometric consistency while keeping supervision efficient, we reconstruct masked tubelets by coordinate-constrained trilinear reprojection from the latent volume. Finally, we adopt hierarchical masking that combines intra-plane tubelet masking with inter-plane plane dropping, turning reconstruction into cross-plane completion and strengthening global 3D reasoning. Experiments show that coherent 3D anatomy can be decoded from the learned latent volume using a lightweight linear probe, and that masked-plane reconstruction recovers missing views from remaining planes. Moreover, our method improves downstream multi-plane segmentation over state-of-the-art methods, while maintaining favorable efficiency under the same input setting. Code is available at this URL.
-
Journal articleKirolos S, Chehrazi M, Emsley R, et al., 2026, , Archives of Disease in Childhood, ISSN: 0003-9888
-
Journal articleWu Y, Rui H, Wang F, et al., 2026, , IEEE Trans Med Imaging, Vol: PP
Multi-contrast super-resolution (MCSR) is crucial for enhancing MRI but current deep learning methods are limited. They typically require large, paired low- and high-resolution (LR/HR) training datasets, which are scarce, and are trained for fixed upsampling scales. While recent self-supervised methods remove the paired data requirement, they fail to leverage valuable population-level priors. In this work, we propose a novel, decoupled MCSR framework that resolves both limitations. We reformulate MCSR into two stages: (1) an unpaired cross-modal synthesis (uCMS) module, trained once on unpaired population data to learn a robust anatomical prior; and (2) a lightweight, patient-specific implicit re-representation (IrR) module. This IrR module is optimized in a self-supervised manner to fuse the population prior with the subject's own LR target data. This design uniquely fuses population-level knowledge with patient-specific fidelity without requiring paired target-domain LR/HR training data or paired cross-modal HR reference-target training data. Here, 'unpaired' refers to the population-level training setting; at subject-specific inference, as in standard MCSR, the method uses a matched HR reference image from another contrast together with the subject's LR target image. By building the IrR module on an implicit neural representation, our framework is also inherently scale-agnostic. Our method demonstrates superior quantitative performance on different datasets, with exceptional robustness at extreme scales (16×, 32×), a regime where competing methods fail. Our work presents a data-efficient, flexible, and computationally lightweight paradigm for MCSR, enabling high-fidelity, arbitrary-scale reconstruction without the need for paired population-level supervision.
-
Journal articleLou J, Pan Q, Yang Q, et al., 2026, , IEEE Transactions on Circuits and Systems for Video Technology, Vol: 36, Pages: 15187-15205, ISSN: 1051-8215
-
Journal articleStatton B, Armour C, Leong K, et al., 2026, , Pulm Circ, Vol: 16, ISSN: 2045-8932
Invasive right heart catheterization (RHC) remains the reference standard for diagnosing pulmonary hypertension (PH), but there is a pressing need for non-invasive techniques to assess pulmonary hemodynamics. This study evaluates 4D-flow MRI-derived wall shear stress (WSS) for PH detection using the updated 2022 hemodynamic criteria, and the prediction of right ventriculo-pulmonary arterial (RV-PA) uncoupling to assess disease severity. 3D WSS was measured in the main (MPA), right (RPA), and left (LPA) pulmonary arteries in a retrospective validation cohort (41 PH patients, 23 non-PH) and then assessed in a prospective test cohort (23 PH patients, 17 healthy controls). For methodological validation, a patient-specific computational fluid dynamics (CFD) was performed in a subset of 16 patients from the validation cohort. RV-PA uncoupling was assessed on the cine MRI, using RV global longitudinal strain (GLS) < -15.3% and RVSV/RVESV < 1. Both 3D peak and time-averaged WSS were significantly reduced in PH patients compared with healthy controls. Global WSS metrics demonstrated the strongest correlations with invasive hemodynamic parameters, including mPAP (r = -0.75, p < 0.001) and PVR (r = -0.76, p < 0.001). ROC curve analysis identified a peak WSS cutoff of < 0.30 Pa (AUC 0.98) for PH detection, a threshold achieving 90% sensitivity and 100% specificity in the test cohort. Furthermore, a peak WSS threshold of < 0.18 Pa demonstrated excellent discriminative ability for RV-PA uncoupling. These findings support the potential integration of WSS analysis into routine CMR protocols as a non-invasive complement to RHC, showing promise for both diagnosis and a longitudinal biomarker for monitoring PH.
-
Journal articleBukala BR, Vano LJ, Carr R, et al., 2026, , Psychopharmacology (Berl)
RATIONALE: Tardive dyskinesia (TD) is a neurological syndrome of involuntary repetitive movements which results from treatment with antipsychotic medication. The pathoetiology of TD is not well understood but a possible mechanism involves dopaminergic overactivity in the nigrostriatal pathway. If this theory is correct, then levels of neuromelanin (a long-term marker of dopaminergic activity) should be higher in people with TD than those without TD. OBJECTIVES: The aim of the study was to test the hypothesis that neuromelanin levels are higher in patients with TD relative to those without TD. METHODS: Data from 27 participants (TD: n = 13; non-TD n = 14) with a diagnosis of schizophrenia, all taking antipsychotic drugs, were used. Neuromelanin was measured in the midbrain via Magnetic Resonance Imaging (MRI) and compared between the groups. Movement symptoms were measured using the Abnormal Involuntary Movement Scale (AIMS), and antipsychotic dose recorded. RESULTS: MRI-measured neuromelanin levels were significantly higher in patients with TD, as compared to those without (mean = 0.223; t(17.27) = 3.817, p = 0.001; g = 1.386, 95% CI=[0.559, 2.213]). This remained significant after controlling for age, sex, substantia nigra/ventral tegmental area volume and antipsychotic dose (ANCOVA: F(1,20) = 12.08, p = 0.0024; adjusted β = 0.0301, 95% CI [0.0120, 0.0481]). The most pronounced difference was seen in the ventral substantia nigra. CONCLUSIONS: The finding of higher midbrain neuromelanin in patients with TD aligns with the dopamine overactivity hypothesis of TD aetiology. It also supports a neurobiological basis for treatment of TD with drugs that target presynaptic dopamine activity, such as VMAT2 inhibitors. Additionally, it identifies the ventral substantia nigra as a key locus. Future longitudinal studies are needed to delineate if do
-
Journal articleWei X, Liao Y, Zhu J, et al., 2026, , Expert Systems with Applications, Vol: 327, ISSN: 0957-4174
Automatic prostate tumor segmentation in multiparametric magnetic resonance imaging (mpMRI) can support targeted biopsy and therapy planning, but remains challenging due to the small, heterogeneous and low-contrast nature of the lesions. CNN-based encoder-decoder models capture local detail well, but struggle with long-range reasoning. Transformers excel at global interactions, but are computationally expensive. Mamba/state-space models enable efficient global modelling, but often lack boundary-sensitive local refinement. We propose ProMamba: a hybrid CNN-Mamba segmentation framework combining local detail extraction with efficient global dependency modelling within a hierarchical U-Net architecture. Its dual-branch Pro-SSM block incorporates a depthwise dilated convolution module (DDCM) to extract multi-scale boundary cues, as well as a Mamba-based visual state-space (VSS) branch to aggregate multi-directional long-range context. The two branches are fused via channel concatenation and shuffling while preserving the 2D convolutional pathway and avoiding spatial flattening. Pyramid-pooled skip connections further reduce the semantic gap between the encoder and decoder and improve small-lesion delineation. Using five-fold cross-validation at the patient level on Prostate158, PI-CAI 2022 and an ethics-approved private cohort of 110 cases, ProMamba achieves Dice scores of 51.39%, 59.46% and 48.49% respectively, outperforming strong CNN, nnU-Net, efficient Transformer/hybrid and Mamba baselines. ProMamba also shows competitive latency and memory usage and generalizes well to PROMISE12 and ACDC, achieving Dice scores of 91.40% and 92.14%, respectively.
-
Journal articleThangarajah D, Chappell K, Burtt AH, et al., 2026, , Inflamm Bowel Dis
BACKGROUND: Adipose tissue measures are poorly defined in pediatric Crohn's disease (CD). Visceral adipose tissue (VAT) plays a role in regulating innate immunity and obesity-related diseases, contributing to poorer long-term health outcomes. METHODS: Abdominal tissue volumes and liver fat were quantified using magnetic resonance imaging and magnetic resonance spectrometry in children with CD. Dietary, adipokine, and metabolic profiles were analyzed in subjects with CD. Multivariable linear regression analysis was performed to evaluate the factors associated with tissue volumes. RESULTS: We recruited 25 children with CD and 8 healthy controls. Children with CD exhibited significantly higher VAT and subcutaneous adipose tissue (SCAT) volumes than their healthy counterparts, with increases of 0.33 L (95% CI, 0.12-0.52, P < .01) and 1.83 L (95% CI, 1.34-2.45, P < .01), respectively. Muscle measures were lower in children with CD, while liver fat measurements and BMI-z scores remained within healthy reference ranges. Disease duration, male sex, and C-reactive protein (CRP) levels were significant predictors of VAT in children with CD. SCAT was associated with plasma leptin levels, and VAT correlated with plasma triglyceride levels. Children with CD had diets low in calories and high in non-milk extrinsic sugars. CONCLUSION: This study describes a specific adiposity pattern in children with Crohn's disease; 'Thin on Outside Fat on the Inside Crohn's associated Obesity' (TOFICO). This is characterized by increased visceral and subcutaneous adipose tissue in the context of normal liver fat and BMI. We have shown that visceral obesity is a robust marker of bowel inflammation in Crohn's disease and is not affected by anthropometric factors, unlike BMI.
-
Journal articleWang A, Lei M, Zhang H, et al., 2026, , IEEE Transactions on Medical Imaging, Vol: 45, Pages: 4714-4727, ISSN: 0278-0062
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.
Contact
For enquiries about the MRI Physics Collective, please contact:
Mary Finnegan
Senior MR Physicist at the ImperiaCollege Healthcare NHS Trust
Pete Lally
Assistant Professor in Magnetic Resonance (MR) Physics at ImperiaCollege
Jan Sedlacik
MR Physicist at the Robert Steiner MR Unit, Hammersmith Hospital Campus