Medical image analysis最新文献

{"title":"SupReMix: Supervised contrastive learning for medical imaging regression with mixup","authors":"Yilei Wu ,&nbsp;Zijian Dong ,&nbsp;Chongyao Chen ,&nbsp;Wangchunshu Zhou ,&nbsp;Juan Helen Zhou","doi":"10.1016/j.media.2025.103909","DOIUrl":"10.1016/j.media.2025.103909","url":null,"abstract":"<div><div>In medical image analysis, regression plays a critical role in computer-aided diagnosis. It enables quantitative measurements such as age prediction from structural imaging, cardiac function quantification, and molecular measurement from PET scans. While deep learning has shown promise for these tasks, most approaches focus solely on optimizing regression loss or model architecture, neglecting the quality of learned feature representations which are crucial for robust clinical predictions. Directly applying representation learning techniques designed for classification to regression often results in fragmented representations in the latent space, yielding sub-optimal performance. In this paper, we argue that the potential of contrastive learning for medical image regression has been overshadowed due to the neglect of two crucial aspects: <em>ordinality-awareness</em> and <em>hardness</em>. To address these challenges, we propose <strong>Sup</strong>ervised Contrastive Learning for Medical Imaging <strong>Re</strong>gression with <strong>Mi</strong>xup (<strong>SupReMix</strong>). It takes anchor-inclusive mixtures (mixup of the anchor and a distinct negative sample) as hard negative pairs and anchor-exclusive mixtures (mixup of two distinct negative samples) as hard positive pairs at the embedding level. This strategy formulates harder contrastive pairs by integrating richer ordinal information. Through theoretical analysis and extensive experiments on six datasets spanning MRI, X-ray, ultrasound, and PET modalities, we demonstrate that SupReMix fosters continuous ordered representations, significantly improving regression performance.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103909"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145894000","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}

{"title":"Unsupervised anomaly detection in brain MRI via disentangled anatomy learning","authors":"Tao Yang ,&nbsp;Xiuying Wang ,&nbsp;Hao Liu ,&nbsp;Guanzhong Gong ,&nbsp;Lian-Ming Wu ,&nbsp;Yu-Ping Wang ,&nbsp;Lisheng Wang","doi":"10.1016/j.media.2025.103922","DOIUrl":"10.1016/j.media.2025.103922","url":null,"abstract":"<div><div>Detection of various lesions in brain MRI is clinically critical, but challenging due to the diversity of lesions and variability in imaging conditions. Current unsupervised learning methods detect anomalies mainly through reconstructing abnormal images into pseudo-healthy images (PHIs) by normal samples learning and then analyzing differences between images. However, these unsupervised models face two significant limitations: restricted generalizability to multi-modality and multi-center MRIs due to their reliance on the specific imaging information in normal training data, and constrained performance due to abnormal residuals propagated from input images to reconstructed PHIs. To address these limitations, two novel modules are proposed, forming a new PHI reconstruction framework. Firstly, the disentangled representation module is proposed to improve generalizability by decoupling brain MRI into imaging information and essential imaging-invariant anatomical images, ensuring that the reconstruction focuses on the anatomy. Specifically, brain anatomical priors and a differentiable one-hot encoding operator are introduced to constrain the disentanglement results and enhance the disentanglement stability. Secondly, the edge-to-image restoration module is designed to reconstruct high-quality PHIs by restoring the anatomical representation from the high-frequency edge information of anatomical images, and then recoupling the disentangled imaging information. This module not only suppresses abnormal residuals in PHI by reducing abnormal pixels input through edge-only input, but also effectively reconstructs normal regions using the preserved structural details in the edges. Evaluated on nine public datasets (4,443 patients’ MRIs from multiple centers), our method outperforms 17 state-of-the-art methods, achieving absolute improvements of +18.32 % in average precision and +13.64 % in Dice similarity coefficient.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103922"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145822817","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}

{"title":"SAM-Swin: SAM-driven dual-swin transformers with adaptive lesion enhancement for Laryngo-Pharyngeal tumor detection","authors":"Jia Wei ,&nbsp;Yun Li ,&nbsp;Xiaomao Fan ,&nbsp;Wenjun Ma ,&nbsp;Meiyu Qiu ,&nbsp;Hongyu Chen ,&nbsp;Wenbin Lei","doi":"10.1016/j.media.2025.103906","DOIUrl":"10.1016/j.media.2025.103906","url":null,"abstract":"<div><div>Laryngo-pharyngeal cancer (LPC) is a highly lethal malignancy in the head and neck region. Recent advancements in tumor detection, particularly through dual-branch network architectures, have significantly improved diagnostic accuracy by integrating global and local feature extraction. However, challenges remain in accurately localizing lesions and fully capitalizing on the complementary nature of features within these branches. To address these issues, we propose SAM-Swin, an innovative SAM-driven Dual-Swin Transformer for laryngo-pharyngeal tumor detection. This model leverages the robust segmentation capabilities of the Segment Anything Model 2 (SAM2) to achieve precise lesion segmentation. Meanwhile, we present a multi-scale lesion-aware enhancement module (MS-LAEM) designed to adaptively enhance the learning of nuanced complementary features across various scales, improving the quality of feature extraction and representation. Furthermore, we implement a multi-scale class-aware guidance (CAG) loss that delivers multi-scale targeted supervision, thereby enhancing the model’s capacity to extract class-specific features. To validate our approach, we compiled three LPC datasets from the First Affiliated Hospital (FAHSYSU), the Sixth Affiliated Hospital (SAHSYSU) of Sun Yat-sen University, and Nanfang Hospital of Southern Medical University (NHSMU). The FAHSYSU dataset is utilized for internal training, while the SAHSYSU and NHSMU datasets serve for external evaluation. Extensive experiments demonstrate that SAM-Swin outperforms state-of-the-art methods, showcasing its potential for advancing LPC detection and improving patient outcomes. The source code of SAM-Swin is available at the URL of <span><span>https://github.com/VVJia/SAM-Swin</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103906"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145705009","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}

{"title":"Complex wavelet-based Transformer for neurodevelopmental disorder diagnosis via direct modeling of real and imaginary components","authors":"Ah-Yeong Jeong ,&nbsp;Da-Woon Heo ,&nbsp;Heung-Il Suk","doi":"10.1016/j.media.2025.103914","DOIUrl":"10.1016/j.media.2025.103914","url":null,"abstract":"<div><div>Resting-state functional magnetic resonance imaging (rs-fMRI) measures intrinsic neural activity, and analyzing its frequency-domain characteristics provides insights into brain dynamics. Owing to these properties, rs-fMRI is widely used to investigate brain disorders such as autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD). Conventional frequency-domain analyses typically rely on the Fourier transform, which lacks flexibility in capturing non-stationary neural signals due to its fixed resolution. Furthermore, these methods primarily utilize only real-valued features, such as the magnitude or phase, derived from complex-valued spectral representations. Consequently, direct modeling of the real and imaginary components, particularly within fMRI analyses, remains largely unexplored, overlooking the distinct and complementary spectral information encoded in these components. To address these limitations, we propose a novel Transformer-based framework that explicitly models the real and imaginary components of continuous wavelet transform (CWT) coefficients from rs-fMRI signals. Our architecture integrates spectral, temporal, and spatial attention modules, employing self- and cross-attention mechanisms to jointly capture intra- and inter-component relationships. Applied to the Autism Brain Imaging Data Exchange (ABIDE)-I and ADHD-200 datasets, our approach achieved state-of-the-art classification performance compared to existing baselines. Comprehensive ablation studies demonstrated the advantages of directly utilizing real and imaginary components over conventional frequency-domain features and validate each module’s contribution. Moreover, attention-based analyses revealed frequency- and region-specific patterns consistent with known neurobiological alterations in ASD and ADHD. These findings highlight that preserving and jointly leveraging the real and imaginary components of CWT-based representations not only enhances diagnostic performance but also provides interpretable insights into neurodevelopmental disorders.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103914"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145753384","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}

{"title":"STCMT-Net: A spatiotemporal consistency motion transfer network for enhancing cardiac motion estimation","authors":"Xiaoya Qiao ,&nbsp;Jiwei Yu ,&nbsp;Hanzhong Wang ,&nbsp;Wenxiang Ding ,&nbsp;Ruiyan Zhang ,&nbsp;Zhengbin Zhu ,&nbsp;Qiu Huang","doi":"10.1016/j.media.2025.103926","DOIUrl":"10.1016/j.media.2025.103926","url":null,"abstract":"<div><div>Cardiac motion estimation is critical for assessing cardiac function and understanding cardiac mechanics. However, the complex and subject-specific characteristics of cardiac motion pose substantial challenges for modeling the spatiotemporal dynamics of the heart from four-dimensional (4D) cardiac images. Existing methods often exhibit limited capacity for spatial representation of cardiac anatomy and inadequate modeling of inter-structural correlations across images, resulting in inaccurate motion estimation throughout the cardiac cycle. To address these challenges, we propose an unsupervised Spatiotemporal Consistency Motion Transfer Network (STCMT-Net) to enhance motion estimation in 4D cardiac images. Cardiac motion is explicitly modeled in a set of keypoints, which are unsupervisedly extracted from continuous images via a keypoint detector. The detected keypoints encode essential motion patterns and serve as structural anchors tightly coupled with the dense motion estimation. The keypoint extraction and matching process are regularized by constraints that ensure spatial distribution uniformity and preserve temporal local anatomical correspondence around keypoints across images. For dense motion reconstruction, the cardiac motion field is formulated as a weighted combination of basis motion vectors derived from paired keypoints. A dense motion estimation module predicts the combination coefficients and incorporates residual refinements to capture both global and fine-grained motions with improved accuracy. We validated the proposed method on 4D computed tomography and cine magnetic resonance imaging datasets. Cardiac strain analysis further demonstrated the effectiveness of STCMT-Net in distinguishing patients under different cardiac conditions, highlighting its potential for clinical application.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103926"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145844787","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}

{"title":"Spatial-frequency dual-constrained Mamba diffusion model for cross-modal generation from CFP to FFA","authors":"Qing Liu ,&nbsp;Hongqing Zhu ,&nbsp;Tianwei Qian ,&nbsp;Xun Xu ,&nbsp;Ying Wang ,&nbsp;Bingcang Huang","doi":"10.1016/j.media.2025.103925","DOIUrl":"10.1016/j.media.2025.103925","url":null,"abstract":"<div><div>Fundus Fluorescein Angiography (FFA) is a critical imaging technique for visualizing retinal vascular dynamics and diagnosing various retinal pathologies. However, its reliance on intravenous fluorescein dye may pose risks of nausea, allergies, and even life-threatening complications, with potential adverse reaction hazards. To overcome these limitations, we propose a novel Spatial-Frequency Dual-Constraint Mamba Diffusion Model (SFDC-MambaDiff) for cross-modal generation of FFA images from non-invasive Color Fundus Photography (CFP). Specifically, we design a Learnable Wavelet Frequency-domain Extractor (LWFE) that integrates spiral Mamba scanning to capture frequency-domain vascular features, serving as structural priors. In parallel, a Dual-pyramid Spatial-domain Feature Extractor (DSFE) is developed by combining convolution and average pooling operations with row-column bidirectional scanning to model both global and local spatial representations, which serve as lesion-aware priors. These two domain-specific priors are further integrated through a Dual-domain Conditional Constraint Mamba module (DCCM), which employs a dual attention mechanism to guide the progressive denoising process of the diffusion model. Experimental results on both public and private datasets demonstrate that SFDC-MambaDiff can synthesize high-fidelity, non-invasive FFA images, offering a promising alternative for safe retinal screening and low-cost diagnostic support. The code for this project is available at <span><span>https://github.com/MISlq/SFDC-MambaDiff</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103925"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145844788","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}

{"title":"NeuroDetour: A neural pathway transformer for uncovering structural-functional coupling mechanisms in human connectome","authors":"Ziquan Wei ,&nbsp;Tingting Dan ,&nbsp;Jiaqi Ding ,&nbsp;Paul J. Laurienti ,&nbsp;Guorong Wu","doi":"10.1016/j.media.2025.103931","DOIUrl":"10.1016/j.media.2025.103931","url":null,"abstract":"<div><div>Although modern imaging methods enable in-vivo examination of connections between distinct brain areas, we still lack a comprehensive understanding of how anatomical structure underpins brain function and how spontaneous fluctuations in neural activity give rise to cognition. At the same time, many efforts in machine learning have focused on modeling the complex, nonlinear relationships between neuroimaging signals and observable traits. Yet, current machine learning techniques often overlook fundamental neuroscience insights, making it difficult to interpret transient neural dynamics in terms of cognitive behavior. To bridge this gap, we turn our attention to the interplay between structural connectivity (SC) and functional connectivity (FC), reframing this open question in network neuroscience as a graph representation learning task centered on neural pathways. In particular, we introduce the notion of a “topological detour” to describe how a given instance of FC (i.e., a direct functional connection) is physically supported by underlying SC pathways (the detour), forming a feedback loop between brain structure and function. By considering these multi-hop detour routes that mediate SC-FC coupling, we design a novel multi-head self-attention mechanism within a Transformer architecture. Building on these ideas, we present a biologically inspired deep-learning framework, <em>NeuroDetour</em>, that extracts connectomic feature representations from large-scale neuroimaging datasets and can be applied to downstream tasks such as task classification and disease prediction. We validated <em>NeuroDetour</em> on extensive public cohorts, including the Human Connectome Project (HCP) and UK Biobank (UKB), using both supervised learning and zero-shot settings. In all scenarios, <em>NeuroDetour</em> achieves state-of-the-art results.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103931"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145956954","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}

{"title":"Extreme cardiac MRI analysis under respiratory motion: Results of the CMRxMotion challenge","authors":"Kang Wang ,&nbsp;Chen Qin ,&nbsp;Zhang Shi ,&nbsp;Haoran Wang ,&nbsp;Xiwen Zhang ,&nbsp;Chen Chen ,&nbsp;Cheng Ouyang ,&nbsp;Chengliang Dai ,&nbsp;Yuanhan Mo ,&nbsp;Chenchen Dai ,&nbsp;Xutong Kuang ,&nbsp;Ruizhe Li ,&nbsp;Xin Chen ,&nbsp;Xiuzheng Yue ,&nbsp;Song Tian ,&nbsp;Alejandro Mora-Rubio ,&nbsp;Kumaradevan Punithakumar ,&nbsp;Shizhan Gong ,&nbsp;Qi Dou ,&nbsp;Sina Amirrajab ,&nbsp;Shuo Wang","doi":"10.1016/j.media.2025.103883","DOIUrl":"10.1016/j.media.2025.103883","url":null,"abstract":"<div><div>Deep learning models have achieved state-of-the-art performance in automated Cardiac Magnetic Resonance (CMR) analysis. However, the efficacy of these models is highly dependent on the availability of high-quality, artifact-free images. In clinical practice, CMR acquisitions are frequently degraded by respiratory motion, yet the robustness of deep learning models against such artifacts remains an underexplored problem. To promote research in this domain, we organized the <em>MICCAI CMRxMotion</em> challenge. We curated and publicly released a dataset of 320 CMR cine series from 40 healthy volunteers who performed specific breathing protocols to induce a controlled spectrum of motion artifacts. The challenge comprised two tasks: 1) automated image quality assessment to classify images based on motion severity, and 2) robust myocardial segmentation in the presence of motion artifacts. A total of 22 algorithms were submitted and evaluated on the two designated tasks. This paper presents a comprehensive overview of the challenge design and dataset, reports the evaluation results for the top-performing methods, and further investigates the impact of motion artifacts on five clinically relevant biomarkers. All resources and code are publicly available at: <span><span>https://github.com/CMRxMotion</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103883"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145567299","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}

{"title":"MIRAGE: Medical image-text pre-training for robustness against noisy environments","authors":"Pujin Cheng ,&nbsp;Yijin Huang ,&nbsp;Li Lin ,&nbsp;Junyan Lyu ,&nbsp;Kenneth Kin-Yip Wong ,&nbsp;Xiaoying Tang","doi":"10.1016/j.media.2025.103912","DOIUrl":"10.1016/j.media.2025.103912","url":null,"abstract":"<div><div>Contrastive vision-language pre-training models have achieved significant success on large-scale general multi-modality datasets. However, in the medical domain, the high costs of data collection and expert annotation are likely to result in small-sized and noisy datasets, which can severely limit model performance due to overfitting unreliable data and misrepresenting patterns. To address this challenge, we present MIRAGE, a novel framework designed to handle mismatched false positives and semantically related false negatives during medical image-text pre-training. Cross-entropy-based optimization proves inadequate for noisy contrastive settings, as it tends to fail in distinguishing noisy samples and ends up fitting them, leading to suboptimal representations. To overcome this limitation, we introduce an optimal transport-based contrastive loss that effectively identifies noisy samples leveraging the nearest cross-modality neighbor prior, thereby reducing noisy samples’ adverse impact. Additionally, we propose an adaptive gradient balancing strategy that mitigates the influence of gradients from noisy samples. Extensive experiments demonstrate that MIRAGE achieves superior performance across six tasks and 14 datasets, largely outperforming representative state-of-the-art methods. Furthermore, comprehensive analyses on synthetic noisy data are performed, clearly demonstrating the contribution of each component in MIRAGE.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103912"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145731918","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}

{"title":"UDV-Net: A hybrid CNN and transformer vein segmentation network with vascular prior and spatial awareness","authors":"Bowei Shen ,&nbsp;Xiaoquan Huang ,&nbsp;Yuli Li ,&nbsp;Xinghuan Li ,&nbsp;Lili Ma ,&nbsp;Yonghong Shi ,&nbsp;Shiyao Chen","doi":"10.1016/j.media.2025.103929","DOIUrl":"10.1016/j.media.2025.103929","url":null,"abstract":"<div><div>CNNs handling multi-scale variations and Transformers modeling long-range dependencies are crucial for vascular segmentation. The fusion of these two models effectively combines the multi-scale local features extracted by CNNs with the global information modeled by Transformers, significantly enhancing the accuracy of blood vessel segmentation. However, the powerful model faces challenges when dealing with the gradual formation of extensive collateral vessels in the upper digestive system veins of patients with cirrhotic portal hypertension, leading to numerous false negative and false positive segmentation results. To this end, the paper proposes UDV-Net, a fusion network combining CNN and Transformer with vessel prior and spatial awareness for upper digestive system vein vessel segmentation. Initially, a CNN utilizing an encoding-decoding architecture is employed to create a multi-scale representation of blood vessels from the image. The representation is further refined by the blood vessel attention module at the corresponding scale to address tubular structures, thereby reducing false positive results. Secondly, a Transformer bridge with three-dimensional voxel position encoding is proposed to connect the corresponding encoding-decoding layer, effectively perceiving widely distributed blood vessels with diverse shapes, improving blood vessel connectivity, and avoiding false negative blood vessel results. We collected and annotated abdominal contrast-enhanced CT images of 191 patients with liver cirrhosis, constituting the PHCT dataset. Our method’s validation result on this dataset is state-of-the-art. When evaluated on the publicly available 3D-IRCADb dataset as an unseen external validation set for PHCT, the model demonstrated satisfactory performance. Additionally, our method also achieves the optimal performance on the public MSD hepatic vessel dataset.</div></div>","PeriodicalId":18328,"journal":{"name":"Medical image analysis","volume":"109 ","pages":"Article 103929"},"PeriodicalIF":11.8,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145893999","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}