Pub Date : 2026-01-22DOI: 10.1016/j.media.2026.103956
Jiaju Huang , Xiao Yang , Xinglong Liang , Shaobin Chen , Yue Sun , Greta Sp Mok , Shuo Li , Ying Wang , Tao Tan
Accurate segmentation of breast cancer in PET-CT images is crucial for precise staging, monitoring treatment response, and guiding personalized therapy. However, the small size and dispersed nature of metastatic lesions, coupled with the scarcity of annotated data and heterogeneity between modalities that hinders effective information fusion, make this task challenging. This paper proposes a novel anatomy-guided cross-modal learning framework to address these issues. Our approach first generates organ pseudo-labels through a teacher-student learning paradigm, which serve as anatomical prompts to guide cancer segmentation. We then introduce a self-aligning cross-modal pre-training method that aligns PET and CT features in a shared latent space through masked 3D patch reconstruction, enabling effective cross-modal feature fusion. Finally, we initialize the segmentation network’s encoder with the pre-trained encoder weights, and incorporate organ labels through a Mamba-based prompt encoder and Hypernet-Controlled Cross-Attention mechanism for dynamic anatomical feature extraction and fusion. Notably, our method outperforms eight state-of-the-art methods, including CNN-based, transformer-based, and Mamba-based approaches, on two datasets encompassing primary breast cancer, metastatic breast cancer, and other types of cancer segmentation tasks.
{"title":"Anatomy-guided prompting with cross-modal self-alignment for whole-body PET-CT breast cancer segmentation","authors":"Jiaju Huang , Xiao Yang , Xinglong Liang , Shaobin Chen , Yue Sun , Greta Sp Mok , Shuo Li , Ying Wang , Tao Tan","doi":"10.1016/j.media.2026.103956","DOIUrl":"10.1016/j.media.2026.103956","url":null,"abstract":"<div><div>Accurate segmentation of breast cancer in PET-CT images is crucial for precise staging, monitoring treatment response, and guiding personalized therapy. However, the small size and dispersed nature of metastatic lesions, coupled with the scarcity of annotated data and heterogeneity between modalities that hinders effective information fusion, make this task challenging. This paper proposes a novel anatomy-guided cross-modal learning framework to address these issues. Our approach first generates organ pseudo-labels through a teacher-student learning paradigm, which serve as anatomical prompts to guide cancer segmentation. We then introduce a self-aligning cross-modal pre-training method that aligns PET and CT features in a shared latent space through masked 3D patch reconstruction, enabling effective cross-modal feature fusion. Finally, we initialize the segmentation network’s encoder with the pre-trained encoder weights, and incorporate organ labels through a Mamba-based prompt encoder and Hypernet-Controlled Cross-Attention mechanism for dynamic anatomical feature extraction and fusion. Notably, our method outperforms eight state-of-the-art methods, including CNN-based, transformer-based, and Mamba-based approaches, on two datasets encompassing primary breast cancer, metastatic breast cancer, and other types of cancer segmentation tasks.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"110 ","pages":"Article 103956"},"PeriodicalIF":11.8,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033893","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-20DOI: 10.1016/j.media.2026.103954
Jinwei Zhang , Lianrui Zuo , Blake E. Dewey , Samuel W. Remedios , Yihao Liu , Savannah P. Hays , Dzung L. Pham , Ellen M. Mowry , Scott D. Newsome , Peter A. Calabresi , Shiv Saidha , Aaron Carass , Jerry L. Prince
Automated segmentation of multiple sclerosis (MS) lesions using multicontrast magnetic resonance (MR) images improves efficiency and reproducibility compared to manual delineation, with deep learning (DL) methods achieving state-of-the-art performance. However, these DL-based methods have yet to simultaneously optimize in-domain accuracy and out-of-domain generalization when trained on a single source with limited data, or their performance has been unsatisfactory. To fill this gap, we propose a method called UNISELF, which achieves high accuracy within a single training domain while demonstrating strong generalizability across multiple out-of-domain test datasets. UNISELF employs a novel test-time self-ensembled lesion fusion to improve segmentation accuracy, and leverages test-time instance normalization (TTIN) of latent features to address domain shifts and missing input contrasts. Trained on the ISBI 2015 longitudinal MS segmentation challenge training dataset, UNISELF ranks among the best-performing methods on the challenge test dataset. Additionally, UNISELF outperforms all benchmark methods trained on the same ISBI training data across diverse out-of-domain test datasets with domain shifts and missing contrasts, including the public MICCAI 2016 and UMCL datasets, as well as a private multisite dataset. These test datasets exhibit domain shifts and/or missing contrasts caused by variations in acquisition protocols, scanner types, and imaging artifacts arising from imperfect acquisition. Our code is available at https://github.com/Jinwei1209/UNISELF.
{"title":"UNISELF: A unified network with instance normalization and self-ensembled lesion fusion for multiple sclerosis lesion segmentation","authors":"Jinwei Zhang , Lianrui Zuo , Blake E. Dewey , Samuel W. Remedios , Yihao Liu , Savannah P. Hays , Dzung L. Pham , Ellen M. Mowry , Scott D. Newsome , Peter A. Calabresi , Shiv Saidha , Aaron Carass , Jerry L. Prince","doi":"10.1016/j.media.2026.103954","DOIUrl":"10.1016/j.media.2026.103954","url":null,"abstract":"<div><div>Automated segmentation of multiple sclerosis (MS) lesions using multicontrast magnetic resonance (MR) images improves efficiency and reproducibility compared to manual delineation, with deep learning (DL) methods achieving state-of-the-art performance. However, these DL-based methods have yet to simultaneously optimize in-domain accuracy and out-of-domain generalization when trained on a single source with limited data, or their performance has been unsatisfactory. To fill this gap, we propose a method called UNISELF, which achieves high accuracy within a single training domain while demonstrating strong generalizability across multiple out-of-domain test datasets. UNISELF employs a novel test-time self-ensembled lesion fusion to improve segmentation accuracy, and leverages test-time instance normalization (TTIN) of latent features to address domain shifts and missing input contrasts. Trained on the ISBI 2015 longitudinal MS segmentation challenge training dataset, UNISELF ranks among the best-performing methods on the challenge test dataset. Additionally, UNISELF outperforms all benchmark methods trained on the same ISBI training data across diverse out-of-domain test datasets with domain shifts and missing contrasts, including the public MICCAI 2016 and UMCL datasets, as well as a private multisite dataset. These test datasets exhibit domain shifts and/or missing contrasts caused by variations in acquisition protocols, scanner types, and imaging artifacts arising from imperfect acquisition. Our code is available at <span><span>https://github.com/Jinwei1209/UNISELF</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103954"},"PeriodicalIF":11.8,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-20DOI: 10.1016/j.media.2026.103955
Xin Zhu , Ahmet Enis Cetin , Gorkem Durak , Batuhan Gundogdu , Ziliang Hong , Hongyi Pan , Ertugrul Aktas , Elif Keles , Hatice Savas , Aytekin Oto , Hiten Patel , Adam B. Murphy , Ashley Ross , Frank Miller , Baris Turkbey , Ulas Bagci
Bias field artifacts in magnetic resonance imaging (MRI) scans introduce spatially smooth intensity inhomogeneities that degrade image quality and hinder downstream analysis. To address this challenge, we propose a novel variational Hadamard U-Net (VHU-Net) for effective body MRI bias field correction. The encoder comprises multiple convolutional Hadamard transform blocks (ConvHTBlocks), each integrating convolutional layers with a Hadamard transform (HT) layer. Specifically, the HT layer performs channel-wise frequency decomposition to isolate low-frequency components, while a subsequent scaling layer and semi-soft thresholding mechanism suppress redundant high-frequency noise. To compensate for the HT layer’s inability to model inter-channel dependencies, the decoder incorporates an inverse HT-reconstructed transformer block, enabling global, frequency-aware attention for the recovery of spatially consistent bias fields. The stacked decoder ConvHTBlocks further enhance the capacity to reconstruct the underlying ground-truth bias field. Building on the principles of variational inference, we formulate a new evidence lower bound (ELBO) as the training objective, promoting sparsity in the latent space while ensuring accurate bias field estimation. Comprehensive experiments on body MRI datasets demonstrate the superiority of VHU-Net over existing state-of-the-art methods in terms of intensity uniformity. Moreover, the corrected images yield substantial downstream improvements in segmentation accuracy. Our framework offers computational efficiency, interpretability, and robust performance across multi-center datasets, making it suitable for clinical deployment. The codes are available at https://github.com/Holmes696/Probabilistic-Hadamard-U-Net.
{"title":"VHU-Net: Variational hadamard U-Net for body MRI bias field correction","authors":"Xin Zhu , Ahmet Enis Cetin , Gorkem Durak , Batuhan Gundogdu , Ziliang Hong , Hongyi Pan , Ertugrul Aktas , Elif Keles , Hatice Savas , Aytekin Oto , Hiten Patel , Adam B. Murphy , Ashley Ross , Frank Miller , Baris Turkbey , Ulas Bagci","doi":"10.1016/j.media.2026.103955","DOIUrl":"10.1016/j.media.2026.103955","url":null,"abstract":"<div><div>Bias field artifacts in magnetic resonance imaging (MRI) scans introduce spatially smooth intensity inhomogeneities that degrade image quality and hinder downstream analysis. To address this challenge, we propose a novel variational Hadamard U-Net (VHU-Net) for effective body MRI bias field correction. The encoder comprises multiple convolutional Hadamard transform blocks (ConvHTBlocks), each integrating convolutional layers with a Hadamard transform (HT) layer. Specifically, the HT layer performs channel-wise frequency decomposition to isolate low-frequency components, while a subsequent scaling layer and semi-soft thresholding mechanism suppress redundant high-frequency noise. To compensate for the HT layer’s inability to model inter-channel dependencies, the decoder incorporates an inverse HT-reconstructed transformer block, enabling global, frequency-aware attention for the recovery of spatially consistent bias fields. The stacked decoder ConvHTBlocks further enhance the capacity to reconstruct the underlying ground-truth bias field. Building on the principles of variational inference, we formulate a new evidence lower bound (ELBO) as the training objective, promoting sparsity in the latent space while ensuring accurate bias field estimation. Comprehensive experiments on body MRI datasets demonstrate the superiority of VHU-Net over existing state-of-the-art methods in terms of intensity uniformity. Moreover, the corrected images yield substantial downstream improvements in segmentation accuracy. Our framework offers computational efficiency, interpretability, and robust performance across multi-center datasets, making it suitable for clinical deployment. The codes are available at <span><span>https://github.com/Holmes696/Probabilistic-Hadamard-U-Net</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"110 ","pages":"Article 103955"},"PeriodicalIF":11.8,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-18DOI: 10.1016/j.media.2026.103952
Xiumei Chen , Xinyue Zhang , Wei Xiong , Tao Wang , Aiwei Jia , Qianjin Feng , Meiyan Huang
Performing genome-wide association analysis (GWAS) between hippocampus and whole-genome data can facilitate disease-related biomarker detection of Alzheimer’s disease (AD). However, most existing studies have prioritized hippocampal volume changes and ignored the morphological variations and subfield differences of the hippocampus in AD progression. This disregard restricts the comprehensive understanding of the associations between hippocampus and whole-genome data, which may result in some potentially specific biomarkers of AD being missed. Moreover, the representation of the complex associations between ultra-high-dimensional imaging and whole-genome data remains an unresolved problem in GWAS. To address these issues, we propose an end-to-end hippocampal surface morphological variation-based genome-wide association analysis network (HSM-GWAS) to explore the nonlinear associations between hippocampal surface morphological variations and whole-genome data for AD-related biomarker detection. First, a multi-modality feature extraction module that includes a graph convolution network and an improved diet network is presented to extract imaging and genetic features from non-Euclidean hippocampal surface and whole-genome data, respectively. Second, a dual contrastive learning-based association analysis module is introduced to map and align genetic features to imaging features, thus narrowing the gap between these features and helping explore the complex associations between hippocampal and whole-genome data. Last, a dual cross-attention fusion module is applied to combine imaging and genetic features for disease diagnosis and biomarker detection of AD. Extensive experiments on the real Alzheimer’s Disease Neuroimaging Initiative dataset and simulated data demonstrate that HSM-GWAS considerably improves biomarker detection and disease diagnosis. These findings highlight the ability of HSM-GWAS to discover disease-related biomarkers, suggesting its potential to provide new insights into pathological mechanisms and aid in AD diagnosis. The codes are to be made publicly available at https://github.com/Meiyan88/HSM-GWAS.
{"title":"Hippocampal surface morphological variation-based genome-wide association analysis network for biomarker detection of Alzheimer’s disease","authors":"Xiumei Chen , Xinyue Zhang , Wei Xiong , Tao Wang , Aiwei Jia , Qianjin Feng , Meiyan Huang","doi":"10.1016/j.media.2026.103952","DOIUrl":"10.1016/j.media.2026.103952","url":null,"abstract":"<div><div>Performing genome-wide association analysis (GWAS) between hippocampus and whole-genome data can facilitate disease-related biomarker detection of Alzheimer’s disease (AD). However, most existing studies have prioritized hippocampal volume changes and ignored the morphological variations and subfield differences of the hippocampus in AD progression. This disregard restricts the comprehensive understanding of the associations between hippocampus and whole-genome data, which may result in some potentially specific biomarkers of AD being missed. Moreover, the representation of the complex associations between ultra-high-dimensional imaging and whole-genome data remains an unresolved problem in GWAS. To address these issues, we propose an end-to-end hippocampal surface morphological variation-based genome-wide association analysis network (HSM-GWAS) to explore the nonlinear associations between hippocampal surface morphological variations and whole-genome data for AD-related biomarker detection. First, a multi-modality feature extraction module that includes a graph convolution network and an improved diet network is presented to extract imaging and genetic features from non-Euclidean hippocampal surface and whole-genome data, respectively. Second, a dual contrastive learning-based association analysis module is introduced to map and align genetic features to imaging features, thus narrowing the gap between these features and helping explore the complex associations between hippocampal and whole-genome data. Last, a dual cross-attention fusion module is applied to combine imaging and genetic features for disease diagnosis and biomarker detection of AD. Extensive experiments on the real Alzheimer’s Disease Neuroimaging Initiative dataset and simulated data demonstrate that HSM-GWAS considerably improves biomarker detection and disease diagnosis. These findings highlight the ability of HSM-GWAS to discover disease-related biomarkers, suggesting its potential to provide new insights into pathological mechanisms and aid in AD diagnosis. The codes are to be made publicly available at <span><span>https://github.com/Meiyan88/HSM-GWAS</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"110 ","pages":"Article 103952"},"PeriodicalIF":11.8,"publicationDate":"2026-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995246","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-17DOI: 10.1016/j.media.2026.103949
Kai Gao , Lubin Wang , Liang Li , Xiao Chen , Bin Lu , Yu-Wei Wang , Xue-Ying Li , Zi-Han Wang , Hui-Xian Li , Yi-Fan Liao , Li-Ping Cao , Guan-Mao Chen , Jian-Shan Chen , Tao Chen , Tao-Lin Chen , Yan-Rong Chen , Yu-Qi Cheng , Zhao-Song Chu , Shi-Xian Cui , Xi-Long Cui , Dewen Hu
Functional magnetic resonance imaging (fMRI) allows real-time observation of brain activity through blood oxygen level-dependent (BOLD) signals and is extensively used in studies related to sex classification, age estimation, behavioral measurements prediction, and mental disorder diagnosis. However, the application of deep learning techniques to brain fMRI analysis is hindered by the small sample size of fMRI datasets. Transfer learning offers a solution to this problem, but most existing approaches are designed for large-scale 2D natural images. The heterogeneity between 4D fMRI data and 2D natural images makes direct model transfer infeasible. This study proposes a novel geometric mapping-based fMRI transfer learning method that enables transfer learning from 2D natural images to 4D fMRI brain images, bridging the transfer learning gap between fMRI data and natural images. The proposed Multi-scale Multi-domain Feature Aggregation (MMFA) module extracts effective aggregated features and reduces the dimensionality of fMRI data to 3D space. By treating the cerebral cortex as a folded Riemannian manifold in 3D space and mapping it into 2D space using surface geometric mapping, we make the transfer learning from 2D natural images to 4D brain images possible. Moreover, the topological relationships of the cerebral cortex are maintained with our method, and calculations are performed along the Riemannian manifold of the brain, effectively addressing signal interference problems. The experimental results based on the Human Connectome Project (HCP) dataset demonstrate the effectiveness of the proposed method. Our method achieved state-of-the-art performance in sex classification, age estimation, and behavioral measurement prediction tasks. Moreover, we propose a cascaded transfer learning approach for depression diagnosis, and proved its effectiveness on 23 depression datasets. In summary, the proposed fMRI transfer learning method, which accounts for the structural characteristics of the brain, is promising for applying transfer learning from natural images to brain fMRI images, significantly enhancing the performance in various fMRI analysis tasks.
{"title":"Transfer learning from 2D natural images to 4D fMRI brain images via geometric mapping","authors":"Kai Gao , Lubin Wang , Liang Li , Xiao Chen , Bin Lu , Yu-Wei Wang , Xue-Ying Li , Zi-Han Wang , Hui-Xian Li , Yi-Fan Liao , Li-Ping Cao , Guan-Mao Chen , Jian-Shan Chen , Tao Chen , Tao-Lin Chen , Yan-Rong Chen , Yu-Qi Cheng , Zhao-Song Chu , Shi-Xian Cui , Xi-Long Cui , Dewen Hu","doi":"10.1016/j.media.2026.103949","DOIUrl":"10.1016/j.media.2026.103949","url":null,"abstract":"<div><div>Functional magnetic resonance imaging (fMRI) allows real-time observation of brain activity through blood oxygen level-dependent (BOLD) signals and is extensively used in studies related to sex classification, age estimation, behavioral measurements prediction, and mental disorder diagnosis. However, the application of deep learning techniques to brain fMRI analysis is hindered by the small sample size of fMRI datasets. Transfer learning offers a solution to this problem, but most existing approaches are designed for large-scale 2D natural images. The heterogeneity between 4D fMRI data and 2D natural images makes direct model transfer infeasible. This study proposes a novel geometric mapping-based fMRI transfer learning method that enables transfer learning from 2D natural images to 4D fMRI brain images, bridging the transfer learning gap between fMRI data and natural images. The proposed Multi-scale Multi-domain Feature Aggregation (MMFA) module extracts effective aggregated features and reduces the dimensionality of fMRI data to 3D space. By treating the cerebral cortex as a folded Riemannian manifold in 3D space and mapping it into 2D space using surface geometric mapping, we make the transfer learning from 2D natural images to 4D brain images possible. Moreover, the topological relationships of the cerebral cortex are maintained with our method, and calculations are performed along the Riemannian manifold of the brain, effectively addressing signal interference problems. The experimental results based on the Human Connectome Project (HCP) dataset demonstrate the effectiveness of the proposed method. Our method achieved state-of-the-art performance in sex classification, age estimation, and behavioral measurement prediction tasks. Moreover, we propose a cascaded transfer learning approach for depression diagnosis, and proved its effectiveness on 23 depression datasets. In summary, the proposed fMRI transfer learning method, which accounts for the structural characteristics of the brain, is promising for applying transfer learning from natural images to brain fMRI images, significantly enhancing the performance in various fMRI analysis tasks.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"110 ","pages":"Article 103949"},"PeriodicalIF":11.8,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-17DOI: 10.1016/j.media.2026.103946
Jasmine Nguyen-Duc , Malte Brammerloh , Melina Cherchali , Inès De Riedmatten , Jean-Baptiste Pérot , Jonathan Rafael-Patiño , Ileana O. Jelescu
Monte Carlo diffusion simulations in numerical substrates are valuable for exploring the sensitivity and specificity of the diffusion MRI (dMRI) signal to realistic cell microstructure features. A crucial component of such simulations is the use of numerical phantoms that accurately represent the target tissue, which is in this case, cerebral white matter (WM). This study introduces CATERPillar (Computational Axonal Threading Engine for Realistic Proliferation), a novel method that simulates the mechanics of axonal growth using overlapping spheres as elementary units. CATERPillar facilitates parallel axon development while preventing collisions, offering user control over key structural parameters such as cellular density, undulation, beading and myelination. Its uniqueness lies in its ability to generate not only realistic axonal structures but also realistic glial cells, enhancing the biological fidelity of simulations. We showed that our grown substrates feature distributions of key morphological parameters that agree with those from histological studies. The structural realism of the astrocytic components was quantitatively validated using Sholl analysis. Furthermore, the time-dependent diffusion in the extra- and intra-axonal compartments accurately reflected expected characteristics of short-range disorder, as predicted by theoretical models. CATERPillar is open source and can be used to (a) develop new acquisition schemes that sensitise the MRI signal to unique tissue microstructure features, (b) test the accuracy of a broad range of analytical models, and (c) build a set of substrates to train machine learning models on.
在数值基底上进行蒙特卡罗扩散模拟对于探索扩散核磁共振成像（dMRI）信号对真实细胞微观结构特征的敏感性和特异性具有重要意义。这种模拟的一个关键组成部分是使用数字幻象来准确地代表目标组织，在这种情况下，就是脑白质（WM）。本文介绍了一种以重叠球体为基本单元模拟轴突生长机制的新方法——毛毛虫（Computational Axonal Threading Engine for Realistic Proliferation）。卡特彼勒促进平行轴突的发育，同时防止碰撞，为用户提供对关键结构参数的控制，如细胞密度、波动、串珠和髓鞘形成。它的独特之处在于它不仅能够生成真实的轴突结构，而且能够生成真实的胶质细胞，从而提高了模拟的生物保真度。我们发现，我们培养的底物具有与组织学研究一致的关键形态参数分布。星形细胞成分的结构真实性被定量验证使用肖尔分析。此外，轴突外室和轴突内室的时间依赖性扩散准确地反映了理论模型预测的短期紊乱的预期特征。卡特彼勒是开源的，可用于(a)开发新的采集方案，使MRI信号对独特的组织微观结构特征敏感，(b)测试广泛分析模型的准确性，以及(c)构建一组基板来训练机器学习模型。
{"title":"CATERPillar: a flexible framework for generating white matter numerical substrates with incorporated glial cells","authors":"Jasmine Nguyen-Duc , Malte Brammerloh , Melina Cherchali , Inès De Riedmatten , Jean-Baptiste Pérot , Jonathan Rafael-Patiño , Ileana O. Jelescu","doi":"10.1016/j.media.2026.103946","DOIUrl":"10.1016/j.media.2026.103946","url":null,"abstract":"<div><div>Monte Carlo diffusion simulations in numerical substrates are valuable for exploring the sensitivity and specificity of the diffusion MRI (dMRI) signal to realistic cell microstructure features. A crucial component of such simulations is the use of numerical phantoms that accurately represent the target tissue, which is in this case, cerebral white matter (WM). This study introduces CATERPillar (Computational Axonal Threading Engine for Realistic Proliferation), a novel method that simulates the mechanics of axonal growth using overlapping spheres as elementary units. CATERPillar facilitates parallel axon development while preventing collisions, offering user control over key structural parameters such as cellular density, undulation, beading and myelination. Its uniqueness lies in its ability to generate not only realistic axonal structures but also realistic glial cells, enhancing the biological fidelity of simulations. We showed that our grown substrates feature distributions of key morphological parameters that agree with those from histological studies. The structural realism of the astrocytic components was quantitatively validated using Sholl analysis. Furthermore, the time-dependent diffusion in the extra- and intra-axonal compartments accurately reflected expected characteristics of short-range disorder, as predicted by theoretical models. CATERPillar is open source and can be used to (a) develop new acquisition schemes that sensitise the MRI signal to unique tissue microstructure features, (b) test the accuracy of a broad range of analytical models, and (c) build a set of substrates to train machine learning models on.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"110 ","pages":"Article 103946"},"PeriodicalIF":11.8,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-17DOI: 10.1016/j.media.2026.103939
Qi Zhang , Xiuyuan Chen , Ziyi He , Kun Wang , Lianming Wu , Hongxing Shen , Jianqi Sun
T2 hyperintensities in spinal cord MR images are crucial biomarkers for conditions such as degenerative cervical myelopathy (DCM). However, current clinical diagnoses primarily rely on manual evaluation. Deep learning methods have shown promise in lesion detection, but most supervised approaches are heavily dependent on large, annotated datasets. Unsupervised anomaly detection (UAD) offers a compelling alternative by eliminating the need for abnormal data annotations. However, existing UAD methods face challenges of domain shifts and task conflict. We propose an Uncertainty-based Unsupervised Anomaly Detection framework, termed U2AD, to address these limitations. Unlike traditional methods, U2AD is designed to be trained and tested within the same clinical dataset, following a “mask-and-reconstruction” paradigm built on a Vision Transformer-based architecture. We introduce an uncertainty-guided masking strategy to resolve task conflicts between normal reconstruction and anomaly detection to achieve an optimal balance. Specifically, we employ a Monte-Carlo inference technique to estimate reconstruction uncertainty mappings during training. By iteratively optimizing reconstruction training under the guidance of both epistemic and aleatoric uncertainty, U2AD improves the normal representation learning while maintaining the sensitivity to anomalies. Experimental results demonstrate that U2AD outperforms existing UAD methods in patient-level identification and segment-level localization of spinal cord T2 hyperintensities. This framework establishes a new benchmark for incorporating uncertainty guidance into UAD. Our code is available at: https://github.com/zhibaishouheilab/U2AD
{"title":"U2AD: Uncertainty-based unsupervised anomaly detection framework for detecting T2 hyperintensity in MRI spinal cord","authors":"Qi Zhang , Xiuyuan Chen , Ziyi He , Kun Wang , Lianming Wu , Hongxing Shen , Jianqi Sun","doi":"10.1016/j.media.2026.103939","DOIUrl":"10.1016/j.media.2026.103939","url":null,"abstract":"<div><div>T2 hyperintensities in spinal cord MR images are crucial biomarkers for conditions such as degenerative cervical myelopathy (DCM). However, current clinical diagnoses primarily rely on manual evaluation. Deep learning methods have shown promise in lesion detection, but most supervised approaches are heavily dependent on large, annotated datasets. Unsupervised anomaly detection (UAD) offers a compelling alternative by eliminating the need for abnormal data annotations. However, existing UAD methods face challenges of domain shifts and task conflict. We propose an <strong>U</strong>ncertainty-based <strong>U</strong>nsupervised <strong>A</strong>nomaly <strong>D</strong>etection framework, termed <em>U</em><sup>2</sup><em>AD</em>, to address these limitations. Unlike traditional methods, <em>U</em><sup>2</sup><em>AD</em> is designed to be trained and tested within the same clinical dataset, following a “mask-and-reconstruction” paradigm built on a Vision Transformer-based architecture. We introduce an uncertainty-guided masking strategy to resolve task conflicts between normal reconstruction and anomaly detection to achieve an optimal balance. Specifically, we employ a Monte-Carlo inference technique to estimate reconstruction uncertainty mappings during training. By iteratively optimizing reconstruction training under the guidance of both epistemic and aleatoric uncertainty, <em>U</em><sup>2</sup><em>AD</em> improves the normal representation learning while maintaining the sensitivity to anomalies. Experimental results demonstrate that <em>U</em><sup>2</sup><em>AD</em> outperforms existing UAD methods in patient-level identification and segment-level localization of spinal cord T2 hyperintensities. This framework establishes a new benchmark for incorporating uncertainty guidance into UAD. Our code is available at: <span><span>https://github.com/zhibaishouheilab/U2AD</span><svg><path></path></svg></span></div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103939"},"PeriodicalIF":11.8,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-16DOI: 10.1016/j.media.2026.103941
Vladyslav Zalevskyi , Thomas Sanchez , Misha Kaandorp , Margaux Roulet , Diego Fajardo-Rojas , Liu Li , Jana Hutter , Hongwei Bran Li , Matthew J. Barkovich , Hui Ji , Luca Wilhelmi , Aline Dändliker , Céline Steger , Mériam Koob , Yvan Gomez , Anton Jakovčić , Melita Klaić , Ana Adžić , Pavel Marković , Gracia Grabarić , Meritxell Bach Cuadra
Accurate fetal brain tissue segmentation and biometric measurement are essential for monitoring neurodevelopment and detecting abnormalities in utero. The Fetal Tissue Annotation (FeTA) Challenges have established robust multi-center benchmarks for evaluating state-of-the-art segmentation methods. This paper presents the results of the 2024 challenge edition, which introduced three key innovations.
First, we introduced a topology-aware metric based on the Euler characteristic difference (ED) to overcome the performance plateau observed with traditional metrics like Dice or Hausdorff distance (HD), as the performance of the best models in segmentation surpassed the inter-rater variability. While the best teams reached similar scores in Dice (0.81-0.82) and HD95 (2.1–2.3 mm), ED provided greater discriminative power: the winning method achieved an ED of 20.9, representing roughly a 50% improvement over the second- and third-ranked teams despite comparable Dice scores.
Second, we introduced a new 0.55T low-field MRI test set, which, when paired with high-quality super-resolution reconstruction, achieved the highest segmentation performance across all test cohorts (Dice=0.86, HD95=1.69, ED=6.26). This provides the first quantitative evidence that low-cost, low-field MRI can match or surpass high-field systems in automated fetal brain segmentation.
Third, the new biometry estimation task exposed a clear performance gap: although the best model reached a mean average percentage error (MAPE) of 7.72%, most submissions failed to outperform a simple gestational-age-based linear regression model (MAPE=9.56%), and all remained above inter-rater variability with a MAPE of 5.38%.
Finally, by analyzing the top-performing models from FeTA 2024 alongside those from previous challenge editions, we identify ensembles of 3D nnU-Net trained on both real and synthetic data with both image- and anatomy-level augmentations as the most effective approaches for fetal brain segmentation. Our quantitative analysis reveals that acquisition site, super-resolution strategy, and image quality are the primary sources of domain shift, informing recommendations to enhance the robustness and generalizability of automated fetal brain analysis methods.
{"title":"Advances in automated fetal brain MRI segmentation and biometry: Insights from the FeTA 2024 challenge","authors":"Vladyslav Zalevskyi , Thomas Sanchez , Misha Kaandorp , Margaux Roulet , Diego Fajardo-Rojas , Liu Li , Jana Hutter , Hongwei Bran Li , Matthew J. Barkovich , Hui Ji , Luca Wilhelmi , Aline Dändliker , Céline Steger , Mériam Koob , Yvan Gomez , Anton Jakovčić , Melita Klaić , Ana Adžić , Pavel Marković , Gracia Grabarić , Meritxell Bach Cuadra","doi":"10.1016/j.media.2026.103941","DOIUrl":"10.1016/j.media.2026.103941","url":null,"abstract":"<div><div>Accurate fetal brain tissue segmentation and biometric measurement are essential for monitoring neurodevelopment and detecting abnormalities in utero. The Fetal Tissue Annotation (FeTA) Challenges have established robust multi-center benchmarks for evaluating state-of-the-art segmentation methods. This paper presents the results of the 2024 challenge edition, which introduced three key innovations.</div><div>First, we introduced a topology-aware metric based on the Euler characteristic difference (ED) to overcome the performance plateau observed with traditional metrics like Dice or Hausdorff distance (HD), as the performance of the best models in segmentation surpassed the inter-rater variability. While the best teams reached similar scores in Dice (0.81-0.82) and HD95 (2.1–2.3 mm), ED provided greater discriminative power: the winning method achieved an ED of 20.9, representing roughly a 50% improvement over the second- and third-ranked teams despite comparable Dice scores.</div><div>Second, we introduced a new 0.55T low-field MRI test set, which, when paired with high-quality super-resolution reconstruction, achieved the highest segmentation performance across all test cohorts (Dice=0.86, HD95=1.69, ED=6.26). This provides the first quantitative evidence that low-cost, low-field MRI can match or surpass high-field systems in automated fetal brain segmentation.</div><div>Third, the new biometry estimation task exposed a clear performance gap: although the best model reached a mean average percentage error (MAPE) of 7.72%, most submissions failed to outperform a simple gestational-age-based linear regression model (MAPE=9.56%), and all remained above inter-rater variability with a MAPE of 5.38%.</div><div>Finally, by analyzing the top-performing models from FeTA 2024 alongside those from previous challenge editions, we identify ensembles of 3D nnU-Net trained on both real and synthetic data with both image- and anatomy-level augmentations as the most effective approaches for fetal brain segmentation. Our quantitative analysis reveals that acquisition site, super-resolution strategy, and image quality are the primary sources of domain shift, informing recommendations to enhance the robustness and generalizability of automated fetal brain analysis methods.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103941"},"PeriodicalIF":11.8,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995249","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Cardiovascular disease (CVD) is one of the leading causes of death and illness across the world. Especially, early prediction of heart failure (HF) is complicated due to the heterogeneity of its clinical presentations and symptoms. These challenges underscore the need for a multidisciplinary approach for comprehensive evaluation of cardiac state. To this end, we specifically select electrocardiogram (ECG) and 3D cardiac anatomy for their complementary coverage of cardiac electrical activities and fine-grained structural modeling. Building upon this, we present a novel pre-training framework, named Anatomy-Electrocardiogram Model (AEM), to explore their complex interactions. AEM adopts a multi-task self-supervised scheme that combines a masked reconstruction objective with a cardiac measurement (CM) regression branch to embed cardiac functional priors and structural details. Unlike image-domain models that typically localize the whole heart within the image, our 3D anatomy is background-free and continuous in 3D space. Hence, the model can naturally concentrate on finer structures at the patch level. The further integration with ECG captures functional dynamics through electrical conduction, encapsulating holistic cardiac representations. Extensive experiments are conducted on the multi-modal datasets collected from the UK Biobank, which contain paired biventricular point cloud anatomy and 12-lead ECG data. Our proposed AEM achieves an area under the receiver operating characteristic curve of 0.8192 for incident HF prediction and a concordance index of 0.6976 for survival prediction under linear evaluation, outperforming the state-of-the-art multi-modal methods. Additionally, we study the interpretability of the disease prediction by observing that our model effectively recognizes clinically plausible patterns and exhibits a high association with clinical features.
{"title":"AEM: An interpretable multi-task multi-modal framework for cardiac disease prediction","authors":"Jiachuan Peng , Marcel Beetz , Abhirup Banerjee , Min Chen , Vicente Grau","doi":"10.1016/j.media.2026.103951","DOIUrl":"10.1016/j.media.2026.103951","url":null,"abstract":"<div><div>Cardiovascular disease (CVD) is one of the leading causes of death and illness across the world. Especially, early prediction of heart failure (HF) is complicated due to the heterogeneity of its clinical presentations and symptoms. These challenges underscore the need for a multidisciplinary approach for comprehensive evaluation of cardiac state. To this end, we specifically select electrocardiogram (ECG) and 3D cardiac anatomy for their complementary coverage of cardiac electrical activities and fine-grained structural modeling. Building upon this, we present a novel pre-training framework, named Anatomy-Electrocardiogram Model (AEM), to explore their complex interactions. AEM adopts a multi-task self-supervised scheme that combines a masked reconstruction objective with a cardiac measurement (CM) regression branch to embed cardiac functional priors and structural details. Unlike image-domain models that typically localize the whole heart within the image, our 3D anatomy is background-free and continuous in 3D space. Hence, the model can naturally concentrate on finer structures at the patch level. The further integration with ECG captures functional dynamics through electrical conduction, encapsulating holistic cardiac representations. Extensive experiments are conducted on the multi-modal datasets collected from the UK Biobank, which contain paired biventricular point cloud anatomy and 12-lead ECG data. Our proposed AEM achieves an area under the receiver operating characteristic curve of 0.8192 for incident HF prediction and a concordance index of 0.6976 for survival prediction under linear evaluation, outperforming the state-of-the-art multi-modal methods. Additionally, we study the interpretability of the disease prediction by observing that our model effectively recognizes clinically plausible patterns and exhibits a high association with clinical features.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103951"},"PeriodicalIF":11.8,"publicationDate":"2026-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号