检索结果-内蒙古大学图书馆

Patch-CNN: Training data-efficient deep learning for high-fidelity diffusion tensor estimation from minimal diffusion protocols

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Goodwin-Allcock, Tobias Gong, Ting Gray, Robert Nachev, Parashkev Zhang, Hui Department of Computer Science Centre for Medical Image Computing University College London London United Kingdom Department of Brain Repair & Rehabilitation Institute of Neurology UCL London United Kingdom

We propose a new method, Patch-CNN, for diffusion tensor (DT) estimation from only six-direction diffusion weighted images (DWI). Deep learning-based methods have been recently proposed for dMRI parameter estimation, using either voxel-wise fully-connected neural networks (FCN) or image-wise convolutional neural networks (CNN). In the acute clinical context—where pressure of time limits the number of imaged directions to a minimum—existing approaches either require an infeasible number of training images volumes (image-wise CNNs), or do not estimate the fibre orientations (voxel-wise FCNs) required for tractogram estimation. To overcome these limitations, we propose Patch-CNN, a neural network with a minimal (non-voxel-wise) convolutional kernel (3×3×3). Compared with voxel-wise FCNs, this has the advantage of allowing the network to leverage local anatomical information. Compared with image-wise CNNs, the minimal kernel vastly reduces training data demand. Evaluated against both conventional model fitting and a voxel-wise FCN, Patch-CNN, trained with a single subject is shown to improve the estimation of both scalar dMRI parameters and fibre orientation from six-direction DWIs. The improved fibre orientation estimation is shown to produce improved tractogram. © 2023, CC BY-NC-SA.

关键词： Diffusion

Quantifying White Matter Hyperintensity and Brain Volumes in Heterogeneous Clinical and Low-Field Portable MRI

学校读者我要写书评

暂无评论

Quantifying White Matter Hyperintensity and Brain Volumes in...

IEEE International Symposium on Biomedical Imaging

作者： Pablo Laso Stefano Cerri Annabel Sorby-Adams Jennifer Guo Farrah Mateen Philipp Goebl Jiaming Wu Peirong Liu Hongwei Bran Li Sean I. Young Benjamin Billot Oula Puonti Gordon Sze Sam Payabavash Adam DeHavenon Kevin N. Sheth Matthew S. Rosen John Kirsch Nicola Strisciuglio Jelmer M. Wolterink Arman Eshaghi Frederik Barkhof W. Taylor Kimberly Juan Eugenio Iglesias Massachusetts General Hospital Harvard Medical School USA Electrical Engineering Mathematics and Computer Science University of Twente The Netherlands Pioneer Centre for AI University of Copenhagen Denmark Centre for Medical Image Computing University College London UK Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology USA Danish Research Centre for Magnetic Resonance Denmark Yale School of Medicine USA

ISBN: (数字)9798350313338

ISBN: (纸本)9798350313345

Brain atrophy and white matter hyperintensity (WMH) are critical neuroimaging features for ascertaining brain injury in cerebrovascular disease and multiple sclerosis. Automated segmentation and quantification is desirable but existing methods require high-resolution MRI with good signal-to-noise ratio (SNR). This precludes application to clinical and low-field portable MRI (pMRI) scans, thus hampering large-scale tracking of atrophy and WMH progression, especially in underserved areas where pMRI has huge potential. Here we present a method that segments white matter hyperintensity and 36 brain regions from scans of any resolution and contrast (including pMRI) without retraining. We show results on eight public datasets and on a private dataset with paired high- and low-field scans (3T and 64mT), where we attain strong correlation between the WMH (ρ=.85) and hippocampal volumes (ρ=.89) estimated at both fields. Our method is publicly available as part of FreeSurfer, at: http://***/fswiki/WMH-SynthSeg.

关键词： Atrophy Neuroimaging image segmentation Multiple sclerosis image resolution Magnetic resonance imaging White matter

Tackling Hallucination from Conditional Models for medical image Reconstruction with DynamicDPS

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kim, Seunghoi Tregidgo, Henry F.J. Figini, Matteo Jin, Chen Joshi, Sarang Alexander, Daniel C. Centre for Medical Image Computing University College London United Kingdom Department of Medical Physics and Biomedical Engineering University College London United Kingdom Department of Computer Science University College London United Kingdom The University of Utah United States Centre for AI DS&AI AstraZeneca United Kingdom

Hallucinations are spurious structures not present in the ground truth, posing a critical challenge in medical image reconstruction, especially for data-driven conditional models. We hypothesize that combining an unconditional diffusion model with data consistency, trained on a diverse dataset, can reduce these hallucinations. Based on this, we propose DynamicDPS, a diffusion-based framework that integrates conditional and unconditional diffusion models to enhance low-quality medical images while systematically reducing hallucinations. Our approach first generates an initial reconstruction using a conditional model, then refines it with an adaptive diffusion-based inverse problem solver. DynamicDPS skips early stage in the reverse process by selecting an optimal starting time point per sample and applies Wolfe’s line search for adaptive step sizes, improving both efficiency and image fidelity. Using diffusion priors and data consistency, our method effectively reduces hallucinations from any conditional model output. We validate its effectiveness in image Quality Transfer for low-field MRI enhancement. Extensive evaluations on synthetic and real MR scans, including a downstream task for tissue volume estimation, show that DynamicDPS reduces hallucinations, improving relative volume estimation by over 15% for critical tissues while using only 5% of the sampling steps required by baseline diffusion models. As a model-agnostic and fine-tuning-free approach, DynamicDPS offers a robust solution for hallucination reduction in medical imaging. The code will be made publicly available upon publication. Copyright © 2025, The Authors. All rights reserved.

关键词： Inverse problems

RAIDER: Rapid, anatomy-independent, deep learning-based PDFF and R*2 estimation using magnitude-only signals

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bray, Timothy J.P. Minore, Giulio V. Bainbridge, Alan Dwyer-Hemmings, Louis Taylor, Stuart A. Hall-Craggs, Margaret A. Zhang, Hui Centre for Medical Imaging University College London London United Kingdom Department of Imaging University College London Hospital London United Kingdom Department of Computer Science Centre for Medical Image Computing University College London London United Kingdom Department of Medical Physics University College London Hospital London United Kingdom

Purpose: There has been substantial recent interest in magnitude-based fitting methods for estimating proton density fat fraction (PDFF) and R*2 from chemical shift-encoded MRI data, since these methods can still be used when complexbased methods fail or when phase data are inaccessible or unreliable, and may also be used as a final processing step with complex-based methods. However, conventional fitting techniques are computationally expensive. Deep learning (DL)-based methods promise to accelerate parameter estimation, but previous attempts have used convolutional neural networks (CNNs), which are limited by training requirements and poor generalizability ('anatomy-dependence'). To address these limitations, we propose RAIDER, a voxelwise method for rapid, anatomy-independent deep learning-based PDFF and R*2 estimation using multiecho magnitude-data. Theory and Methods: RAIDER uses two multilayer perceptrons, each trained separately with simulated single-voxel multi-echo magnitude signals, to estimate PDFF and R*2 . The use of two networks, each with restricted training distribution, solves the problem of degeneracy during training. During inference, the solution from one of the two networks is chosen based on likelihood. Performance and speed are investigated in a series of simulation experiments, in phantoms and in vivo. Results: RAIDER is 285-1450 times faster than conventional magnitude fitting, taking 1.4-2.3s per slice rather than 8-56minutes, and offers performance similar to conventional fitting. It produces accurate PDFF measurements in phantoms and in vivo images with different anatomies, despite having been trained only on simulation data. Conclusion: RAIDER can substantially accelerate magnitude-based PDFF and R*2 estimation, whilst avoiding intrinsic limitations of CNN-based methods. © 2024, CC BY-NC-ND.

关键词： Chemical shift

Survival Analysis for Idiopathic Pulmonary Fibrosis using CT images and Incomplete Clinical Data

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Shahin, Ahmed H. Jacob, Joseph Alexander, Daniel C. Barber, David Department Of Computer Science University College London United Kingdom Centre For Medical Image Computing University College London United Kingdom

Idiopathic Pulmonary Fibrosis (IPF) is an inexorably progressive fibrotic lung disease with a variable and unpredictable rate of progression. CT scans of the lungs inform clinical assessment of IPF patients and contain pertinent information related to disease progression. In this work, we propose a multi-modal method that uses neural networks and memory banks to predict the survival of IPF patients using clinical and imaging data. The majority of clinical IPF patient records have missing data (e.g. missing lung function tests). To this end, we propose a probabilistic model that captures the dependencies between the observed clinical variables and imputes missing ones. This principled approach to missing data imputation can be naturally combined with a deep survival analysis model. We show that the proposed framework yields significantly better survival analysis results than baselines in terms of concordance index and integrated Brier score. Our work also provides insights into novel image-based biomarkers that are linked to mortality. Copyright © 2022, The Authors. All rights reserved.

关键词： Bioinformatics

Choice of training label matters: how to best use deep learning for quantitative MRI parameter estimation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Epstein, Sean C. Bray, Timothy J.P. Hall-Craggs, Margaret Zhang, Hui Department of Computer Science Centre for Medical Image Computing University College London London United Kingdom Centre for Medical Imaging University College London London United Kingdom

Deep learning (DL) is gaining popularity as a parameter estimation method for quantitative MRI. A range of competing implementations have been proposed, relying on either supervised or self-supervised learning. Self-supervised approaches, sometimes referred to as unsupervised, have been loosely based on auto-encoders, whereas supervised methods have, to date, been trained on groundtruth labels. These two learning paradigms have been shown to have distinct strengths. Notably, self-supervised approaches offer lower-bias parameter estimates than their supervised alternatives. This result is counterintuitive – incorporating prior knowledge with supervised labels should, in theory, lead to improved accuracy. In this work, we show that this apparent limitation of supervised approaches stems from the naïve choice of groundtruth training labels. By training on labels which are deliberately not groundtruth, we show that the low-bias parameter estimation previously associated with self-supervised methods can be replicated – and improved on – within a supervised learning framework. This approach sets the stage for a single, unifying, deep learning parameter estimation framework, based on supervised learning, where trade-offs between bias and variance are made by careful adjustment of training label. Copyright © 2022, The Authors. All rights reserved.

关键词： Parameter estimation

AI-assisted prostate cancer detection and localisation on biparametric MR by classifying radiologist-positives

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wu, Xiangcen Wang, Yipei Yang, Qianye Thorley, Natasha Punwani, Shonit Kasivisvanathan, Veeru Bonmati, Ester Hu, Yipeng Centre for Medical Image Computing Department of Medical Physics and Biomedical Engineering University College London London United Kingdom Div of Surgery & Interventional Sci University College London London United Kingdom School of Computer Science and Engineering University of Westminster London United Kingdom

Prostate cancer diagnosis through MR imaging have currently relied on radiologists’ interpretation, whilst modern AI-based methods have been developed to detect clinically significant cancers independent of radiologists. In this study, we propose to develop deep learning models that improve the overall cancer diagnostic accuracy, by classifying radiologist-identified patients or lesions (i.e. radiologist-positives), as opposed to the existing models that are trained to discriminate over all patients. We develop a single voxel-level classification model, with a simple percentage threshold to determine positive cases, at levels of lesions, Barzell-zones and patients. Based on the presented experiments from two clinical data sets, consisting of histopathology-labelled MR images from more than 800 and 500 patients in the respective UCLA and UCL PROMIS studies, we show that the proposed strategy can improve the diagnostic accuracy, by augmenting the radiologist reading of the MR imaging. Among varying definition of clinical significance, the proposed strategy, for example, achieved a specificity of 44.1% (with AI assistance) from 36.3% (by radiologists alone), at a controlled sensitivity of 80.0% on the publicly available UCLA data set. This provides measurable clinical values in a range of applications such as reducing unnecessary biopsies, lowering cost in cancer screening and quantifying risk in therapies. © 2024, CC BY.

关键词： Urology

Opportunities and Barriers for Adoption of a Decision-Support Tool for Alzheimer's Disease

学校读者我要写书评

暂无评论

ACM Transactions on computing for Healthcare 2021年第4期2卷 1–19页

作者： Bellio, Maura Furniss, Dominic Oxtoby, Neil P. Garbarino, Sara Firth, Nicholas C. Ribbens, Annemie Alexander, Daniel C. Blandford, Ann Centre for Medical Image Computing 90 High Holborn Floor 1 LondonWC1V 6LJ United Kingdom UCL Interaction Centre 66-72 Gower St LondonWC1E 6EA United Kingdom MIDA group University of Genova DIMA Room 003 0th floor Via Dodecaneso 35 Genova16146 Italy Icometrix Kolonel Begaultlaan 1b/12 Leuven3012 Belgium Centre for Medical Image Computing Department of Computer Science UCL

Clinical decision-support tools (DSTs) represent a valuable resource in healthcare. However, lack of Human Factors considerations and early design research has often limited their successful adoption. To complement previous technically focused work, we studied adoption opportunities of a future DST built on a predictive model of Alzheimer's Disease (AD) progression. Our aim is two-fold: exploring adoption opportunities for DSTs in AD clinical care, and testing a novel combination of methods to support this process. We focused on understanding current clinical needs and practices, and the potential for such a tool to be integrated into the setting, prior to its development. Our user-centred approach was based on field observations and semi-structured interviews, analysed through workflow analysis, user profiles, and a design-reality gap model. The first two are common practice, whilst the latter provided added value in highlighting specific adoption needs. We identified the likely early adopters of the tool as being both psychiatrists and neurologists based in research-oriented clinical settings. We defined ten key requirements for the translation and adoption of DSTs for AD around IT, user, and contextual factors. Future works can use and build on these requirements to stand a greater chance to get adopted in the clinical setting. © 2021 ACM.

关键词： User centered design

OP0259 WHITE MATTER HYPERINTENSITY VOLUME IN UK BIOBANK PATIENTS WITH RHEUMATOID ARTHRITIS AND ULCERATIVE COLITIS

学校读者我要写书评

暂无评论

Annals of the Rheumatic Diseases 2023年 82卷 171-172页

作者： J. Cox M. De Groot J. Cole S. Williams King's College London Centre for Neuroimaging Sciences Institute of Psychiatry Psychology & Neuroscience London United Kingdom GSK - Stevenage GlaxoSmithKline Research and Development Stevenage United Kingdom University College London Centre for Medical Image Computing Department of Computer Science London United Kingdom

Background Rheumatoid arthritis (RA) and ulcerative colitis (UC) both result in higher levels of systemic inflammation and represent a significantly higher risk of cerebrovascular disease. The presence of this systemic inflammation has been linked to central nervous system effects and both diseases have increased rates of fatigue, brain fog and depression. Additionally, systemic inflammation has been implicated in decreasing the integrity of the blood brain barrier and potentially resulting in further involvement of the central nervous system in autoimmune diseases. Objectives The increased presence of white matter hyperintensities (WMH) has been proposed as a potential link to the central symptoms reported and the increased prevalence of cerebrovascular accidents in these diseases [1]. The present study aimed to analyse total volume of WMH in patients with RA and UC as compared to matched controls. Methods Data from 1,968 individuals in the UK Biobank Imaging sub-study were included in the present study [2]. This was divided into two separate patient populations with separate control groups matched to each in a 1:3 patient:control ratio. Data from 252 individuals (mean age ± SD =65.52 ± 7.03, 71.4% females) were included in the RA patient group with 756 age and sex matched controls (mean age ± SD = 65.52 ± 7.02, 71.4% females). Data from 240 individuals (mean age ± SD = 63.98 ± 7.06, 50.8% females) were included in the UC patient group with 720 age and sex matched controls (mean age ± SD = 63.98 ± 7.05, 50.8% females). This study utilised the total volume of WMH acquired using an automated segmentation approach called BIANCA and provided directly by the UKBiobank. Volumes were taken as a ratio of total brain volume and log transformed prior to a linear model regression being performed. Two separate models were run. One with age, sex, and ICV as covariates and a second model additionally including hypertension. Results We observed significantly larger total WMH volu

关键词：