Artificial Intelligence in Medicine最新文献_第9页

Mining multi-electrode and multi-wave electroencephalogram based time-interval temporal patterns for improved classification capabilities and explainability 挖掘基于多电极和多波脑电图的时间间隔时间模式，以提高分类能力和可解释性

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-09-14 DOI: 10.1016/j.artmed.2025.103269

Ofir Landau, Nir Nissim

Brain-computer interface (BCI) systems, and particularly electroencephalogram (EEG) based BCI systems, have become more widely used in recent years and are utilized in various applications and domains ranging from medicine and marketing to games and entertainment. While different algorithms have been used to analyze EEG data and enable its classification, existing algorithms have two main drawbacks; both their classification and explainability capabilities are limited. Lacking in explainability, they cannot indicate which electrodes and waves led to a classification decision or explain how areas and frequencies of the brain's activity correlate to a specific task.

In this study, we propose a novel extension for the time-interval temporal patterns mining algorithms aimed at enhancing the data mining process by enabling a richer set of patterns to be learned from the EEG data, thereby contributing to improved classification and explainability capabilities. The extended algorithm is designed to capture and leverage the unique nature of EEG data by decomposing it into different brain waves and modeling the relations among them and between different electrodes.

Our evaluation of the proposed extended algorithm on multiple learning tasks and three EEG datasets demonstrated the extended algorithm's ability to mine richer patterns that improve the classification performance by 4–11 % based on the Area-Under the receiver operating characteristic Curve (AUC) metric, compared to the original version of the algorithm. Moreover, the algorithm was shown to shed light on the areas and frequencies of the brain's activity that are correlated with specific tasks.

脑机接口（BCI）系统，特别是基于脑电图（EEG）的脑机接口系统，近年来得到了越来越广泛的应用，并被用于从医学、营销到游戏和娱乐的各种应用和领域。虽然已有不同的算法用于分析脑电数据并实现其分类，但现有算法存在两个主要缺陷；它们的分类和解释能力都是有限的。由于缺乏可解释性，它们无法指出哪些电极和电波导致了分类决策，也无法解释大脑活动的区域和频率如何与特定任务相关联。在这项研究中，我们提出了一种新的时间间隔时间模式挖掘算法的扩展，旨在通过从EEG数据中学习更丰富的模式集来增强数据挖掘过程，从而有助于提高分类和可解释性能力。该扩展算法通过将EEG数据分解为不同的脑电波，并对它们之间以及不同电极之间的关系进行建模，来捕获和利用EEG数据的独特性。我们对提出的扩展算法在多个学习任务和三个脑电图数据集上的评估表明，与原始算法相比，扩展算法能够挖掘更丰富的模式，基于接收者工作特征曲线下的面积（Area-Under the receiver operating characteristic Curve， AUC）指标，将分类性能提高了4 - 11%。此外，该算法还揭示了与特定任务相关的大脑活动的区域和频率。

{"title":"Mining multi-electrode and multi-wave electroencephalogram based time-interval temporal patterns for improved classification capabilities and explainability","authors":"Ofir Landau, Nir Nissim","doi":"10.1016/j.artmed.2025.103269","DOIUrl":"10.1016/j.artmed.2025.103269","url":null,"abstract":"<div><div>Brain-computer interface (BCI) systems, and particularly electroencephalogram (EEG) based BCI systems, have become more widely used in recent years and are utilized in various applications and domains ranging from medicine and marketing to games and entertainment. While different algorithms have been used to analyze EEG data and enable its classification, existing algorithms have two main drawbacks; both their classification and explainability capabilities are limited. Lacking in explainability, they cannot indicate which electrodes and waves led to a classification decision or explain how areas and frequencies of the brain's activity correlate to a specific task.</div><div>In this study, we propose a novel extension for the time-interval temporal patterns mining algorithms aimed at enhancing the data mining process by enabling a richer set of patterns to be learned from the EEG data, thereby contributing to improved classification and explainability capabilities. The extended algorithm is designed to capture and leverage the unique nature of EEG data by decomposing it into different brain waves and modeling the relations among them and between different electrodes.</div><div>Our evaluation of the proposed extended algorithm on multiple learning tasks and three EEG datasets demonstrated the extended algorithm's ability to mine richer patterns that improve the classification performance by 4–11 % based on the Area-Under the receiver operating characteristic Curve (AUC) metric, compared to the original version of the algorithm. Moreover, the algorithm was shown to shed light on the areas and frequencies of the brain's activity that are correlated with specific tasks.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103269"},"PeriodicalIF":6.2,"publicationDate":"2025-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145099517","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Medical multimodal foundation models in clinical diagnosis and treatment: Applications, challenges, and future directions 医学多模态基础模型在临床诊断和治疗中的应用、挑战和未来方向。

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-09-12 DOI: 10.1016/j.artmed.2025.103265

Kai Sun , Siyan Xue , Fuchun Sun , Haoran Sun , Yu Luo , Ling Wang , Siyuan Wang , Na Guo , Lei Liu , Tian Zhao , Xinzhou Wang , Lei Yang , Shuo Jin , Jun Yan , Jiahong Dong

Recent advancements in deep learning have significantly revolutionized the field of clinical diagnosis and treatment, offering novel approaches to improve diagnostic precision and treatment efficacy across diverse clinical domains, thus driving the pursuit of precision medicine. The growing availability of multi-organ and multimodal datasets has accelerated the development of large-scale Medical Multimodal Foundation Models (MMFMs). These models, known for their strong generalization capabilities and rich representational power, are increasingly being adapted to address a wide range of clinical tasks, from early diagnosis to personalized treatment strategies. This review offers a comprehensive analysis of recent developments in MMFMs, focusing on three key aspects: datasets, model architectures, and clinical applications. We also explore the challenges and opportunities in optimizing multimodal representations and discuss how these advancements are shaping the future of healthcare by enabling improved patient outcomes and more efficient clinical workflows.

深度学习的最新进展极大地改变了临床诊断和治疗领域，为提高不同临床领域的诊断精度和治疗效果提供了新的方法，从而推动了对精准医学的追求。越来越多的多器官和多模式数据集的可用性加速了大规模医学多模式基础模型（MMFMs）的发展。这些模型以其强大的泛化能力和丰富的代表性而闻名，越来越多地被用于解决从早期诊断到个性化治疗策略的广泛临床任务。本综述对MMFMs的最新发展进行了全面分析，重点关注三个关键方面：数据集、模型架构和临床应用。我们还探讨了优化多模式表示的挑战和机遇，并讨论了这些进步如何通过改善患者结果和更有效的临床工作流程来塑造医疗保健的未来。

{"title":"Medical multimodal foundation models in clinical diagnosis and treatment: Applications, challenges, and future directions","authors":"Kai Sun , Siyan Xue , Fuchun Sun , Haoran Sun , Yu Luo , Ling Wang , Siyuan Wang , Na Guo , Lei Liu , Tian Zhao , Xinzhou Wang , Lei Yang , Shuo Jin , Jun Yan , Jiahong Dong","doi":"10.1016/j.artmed.2025.103265","DOIUrl":"10.1016/j.artmed.2025.103265","url":null,"abstract":"<div><div>Recent advancements in deep learning have significantly revolutionized the field of clinical diagnosis and treatment, offering novel approaches to improve diagnostic precision and treatment efficacy across diverse clinical domains, thus driving the pursuit of precision medicine. The growing availability of multi-organ and multimodal datasets has accelerated the development of large-scale Medical Multimodal Foundation Models (MMFMs). These models, known for their strong generalization capabilities and rich representational power, are increasingly being adapted to address a wide range of clinical tasks, from early diagnosis to personalized treatment strategies. This review offers a comprehensive analysis of recent developments in MMFMs, focusing on three key aspects: datasets, model architectures, and clinical applications. We also explore the challenges and opportunities in optimizing multimodal representations and discuss how these advancements are shaping the future of healthcare by enabling improved patient outcomes and more efficient clinical workflows.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103265"},"PeriodicalIF":6.2,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145092872","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An end-to-end solution for out-of-hospital emergency medical dispatch triage based on multimodal and continual deep learning 基于多模式和持续深度学习的院外紧急医疗调度分类的端到端解决方案

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-09-10 DOI: 10.1016/j.artmed.2025.103264

Pablo Ferri , Carlos Sáez , Antonio Félix-De Castro , Purificación Sánchez-Cuesta , Juan M. García-Gómez

The objective of this study was to build a multimodal, multitask predictive model—named E2eDeepEMC²—to improve out-of-hospital emergency incident severity assessments while coping with shifts in data distributions over time. We drew on 2 054 694 independent incidents recorded by the Valencian emergency medical dispatch service between 2009 and 2019 (excluding 2013), combining demographic, temporal, clinical and free-text inputs. To handle temporal drift, our model integrates continual learning strategies and comprises three encoder modules (for context, clinical data and text), whose outputs are merged to predict the life-threatening level, admissible response delay and emergency system jurisdiction. Compared with the Valencian Region’s existing in-house triage protocol, E2eDeepEMC² achieved absolute F1-score gains of 18.46% for life-threatening level, 25.96% for response delay and 3.63% for jurisdiction. Compared to non-continual learning baselines, it also outperformed them by 3.04%, 9.66% and 0.58%, respectively. Deployment of E2eDeepEMC² is currently underway in the Valencian Region, underscoring its practical impact on real-world emergency dispatch decision-making.

本研究的目的是建立一个名为e2edeepemc2的多模式、多任务预测模型，以改善院外紧急事件严重程度评估，同时应对数据分布随时间的变化。我们利用了2009年至2019年（不包括2013年）期间瓦伦西亚紧急医疗调度服务记录的2054 694起独立事件，结合了人口统计、时间、临床和自由文本输入。为了处理时间漂移，我们的模型集成了持续学习策略，并包括三个编码器模块（用于上下文，临床数据和文本），其输出被合并以预测危及生命的水平，可接受的响应延迟和紧急系统管辖权。与巴伦西亚地区现有的内部分诊方案相比，E2eDeepEMC2在危及生命水平上的绝对f1评分提高了18.46%，在反应延迟方面提高了25.96%，在管辖权方面提高了3.63%。与非持续学习基线相比，它的表现也分别高出3.04%、9.66%和0.58%。E2eDeepEMC2目前正在巴伦西亚地区部署，强调了其对现实世界应急调度决策的实际影响。

{"title":"An end-to-end solution for out-of-hospital emergency medical dispatch triage based on multimodal and continual deep learning","authors":"Pablo Ferri , Carlos Sáez , Antonio Félix-De Castro , Purificación Sánchez-Cuesta , Juan M. García-Gómez","doi":"10.1016/j.artmed.2025.103264","DOIUrl":"10.1016/j.artmed.2025.103264","url":null,"abstract":"<div><div>The objective of this study was to build a multimodal, multitask predictive model—named E2eDeepEMC<sup>2</sup>—to improve out-of-hospital emergency incident severity assessments while coping with shifts in data distributions over time. We drew on 2054694 independent incidents recorded by the Valencian emergency medical dispatch service between 2009 and 2019 (excluding 2013), combining demographic, temporal, clinical and free-text inputs. To handle temporal drift, our model integrates continual learning strategies and comprises three encoder modules (for context, clinical data and text), whose outputs are merged to predict the life-threatening level, admissible response delay and emergency system jurisdiction. Compared with the Valencian Region’s existing in-house triage protocol, E2eDeepEMC<sup>2</sup> achieved absolute F1-score gains of 18.46% for life-threatening level, 25.96% for response delay and 3.63% for jurisdiction. Compared to non-continual learning baselines, it also outperformed them by 3.04%, 9.66% and 0.58%, respectively. Deployment of E2eDeepEMC<sup>2</sup> is currently underway in the Valencian Region, underscoring its practical impact on real-world emergency dispatch decision-making.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103264"},"PeriodicalIF":6.2,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145050144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

LoRA-PT: Low-rank adapting UNETR for hippocampus segmentation using principal tensor singular values and vectors LoRA-PT：利用主张量奇异值和向量进行海马分割的低秩自适应UNETR

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-09-01 DOI: 10.1016/j.artmed.2025.103254

Guanghua He , Wangang Cheng , Hancan Zhu , Gaohang Yu

The hippocampus is an important brain structure involved in various psychiatric disorders, and its automatic and accurate segmentation is vital for studying these diseases. Recently, deep learning-based methods have made significant progress in hippocampus segmentation. However, training deep neural network models requires substantial computational resources, time, and a large amount of labeled training data, which is frequently scarce in medical image segmentation. To address these issues, we propose LoRA-PT, a novel parameter-efficient fine-tuning (PEFT) method that transfers the pre-trained UNETR model from the BraTS2021 dataset to the hippocampus segmentation task. Specifically, LoRA-PT divides the parameter matrix of the transformer structure into three distinct sizes, yielding three third-order tensors. These tensors are decomposed using tensor singular value decomposition to generate low-rank tensors consisting of the principal singular values and vectors, with the remaining singular values and vectors forming the residual tensor. During fine-tuning, only the low-rank tensors (i.e., the principal tensor singular values and vectors) are updated, while the residual tensors remain unchanged. We validated the proposed method on three public hippocampus datasets, and the experimental results show that LoRA-PT outperformed state-of-the-art PEFT methods in segmentation accuracy while significantly reducing the number of parameter updates. Our source code is available at https://github.com/WangangCheng/LoRA-PT/tree/LoRA-PT.

海马体是涉及多种精神疾病的重要脑结构，其自动准确分割对研究这些疾病至关重要。近年来，基于深度学习的方法在海马体分割方面取得了重大进展。然而，训练深度神经网络模型需要大量的计算资源、时间和大量的标记训练数据，这在医学图像分割中往往是稀缺的。为了解决这些问题，我们提出了LoRA-PT，一种新的参数高效微调（PEFT）方法，将预训练的UNETR模型从BraTS2021数据集转移到海马分割任务中。具体来说，LoRA-PT将变压器结构的参数矩阵分为三个不同的尺寸，产生三个三阶张量。利用张量奇异值分解对这些张量进行分解，得到由主奇异值和向量组成的低秩张量，剩余的奇异值和向量构成残差张量。在微调过程中，只更新低秩张量（即主张量的奇异值和向量），而剩余张量保持不变。我们在三个公开的海马数据集上验证了所提出的方法，实验结果表明，LoRA-PT在分割精度上优于最先进的PEFT方法，同时显著减少了参数更新次数。我们的源代码可从https://github.com/WangangCheng/LoRA-PT/tree/LoRA-PT获得。

{"title":"LoRA-PT: Low-rank adapting UNETR for hippocampus segmentation using principal tensor singular values and vectors","authors":"Guanghua He , Wangang Cheng , Hancan Zhu , Gaohang Yu","doi":"10.1016/j.artmed.2025.103254","DOIUrl":"10.1016/j.artmed.2025.103254","url":null,"abstract":"<div><div>The hippocampus is an important brain structure involved in various psychiatric disorders, and its automatic and accurate segmentation is vital for studying these diseases. Recently, deep learning-based methods have made significant progress in hippocampus segmentation. However, training deep neural network models requires substantial computational resources, time, and a large amount of labeled training data, which is frequently scarce in medical image segmentation. To address these issues, we propose LoRA-PT, a novel parameter-efficient fine-tuning (PEFT) method that transfers the pre-trained UNETR model from the BraTS2021 dataset to the hippocampus segmentation task. Specifically, LoRA-PT divides the parameter matrix of the transformer structure into three distinct sizes, yielding three third-order tensors. These tensors are decomposed using tensor singular value decomposition to generate low-rank tensors consisting of the principal singular values and vectors, with the remaining singular values and vectors forming the residual tensor. During fine-tuning, only the low-rank tensors (i.e., the principal tensor singular values and vectors) are updated, while the residual tensors remain unchanged. We validated the proposed method on three public hippocampus datasets, and the experimental results show that LoRA-PT outperformed state-of-the-art PEFT methods in segmentation accuracy while significantly reducing the number of parameter updates. Our source code is available at <span><span>https://github.com/WangangCheng/LoRA-PT/tree/LoRA-PT</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103254"},"PeriodicalIF":6.2,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144997868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Pathway information on methylation analysis using deep neural network (PROMINENT): An interpretable deep learning method with pathway prior for phenotype prediction using gene-level DNA methylation 使用深度神经网络进行甲基化分析的途径信息：一种可解释的深度学习方法，具有使用基因水平DNA甲基化进行表型预测的途径先验。

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-29 DOI: 10.1016/j.artmed.2025.103236

Soyeon Kim , Laizhi Zhang , Yidi Qin , Rebecca I. Caldino Bohn , Hyun Jung Park

Background

DNA methylation is a key epigenetic marker that influences gene expression and phenotype regulation, and is affected by both genetic and environmental factors. Traditional linear regression methods such as elastic nets have been employed to assess the cumulative effects of multiple DNA methylation markers on phenotypes. However, these methods often fail to capture the complex nonlinear nature of the data. Recent deep learning approaches, such as MethylNet, have improved the prediction accuracy but lack interpretability and efficiency.

Findings

To address these limitations, we introduced Pathway Information on Methylation Analysis using a Deep Neural Network (PROMINENT), a novel interpretable deep learning method that integrates gene-level DNA methylation data with biological pathway information for phenotype prediction. PROMINENT enhances interpretability and prediction accuracy by incorporating gene- and pathway-level priors from databases such as Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). It employs SHapley Additive exPlanations (SHAP) to prioritize significant genes and pathways. Evaluated across various datasets, childhood asthma, idiopathic pulmonary fibrosis (IPF), and first-episode psychosis (FEP)—PROMINENT consistently outperformed existing methods in terms of prediction accuracy and computational efficiency. PROMINENT also identified crucial genes and pathways involved in disease mechanisms.

Conclusions

PROMINENT represents a significant advancement in leveraging DNA methylation data for phenotype prediction, offering both high accuracy and interpretability within reasonable computational time. This method holds promise for elucidating the epigenetic underpinnings of complex diseases and enhancing the utility of DNA methylation data in biomedical research.

背景：DNA甲基化是影响基因表达和表型调控的关键表观遗传标记，受到遗传和环境因素的双重影响。传统的线性回归方法（如弹性网）已被用于评估多个DNA甲基化标记对表型的累积效应。然而，这些方法往往不能捕捉到数据复杂的非线性性质。最近的深度学习方法，如MethylNet，提高了预测的准确性，但缺乏可解释性和效率。为了解决这些局限性，我们引入了使用深度神经网络进行甲基化分析的途径信息，这是一种新的可解释的深度学习方法，将基因水平的DNA甲基化数据与生物学途径信息相结合，用于表型预测。PROMINENT通过整合来自基因本体（GO）和京都基因与基因组百科全书（KEGG）等数据库的基因和途径级先验，提高了可解释性和预测准确性。它采用SHapley加法解释（SHAP）来优先考虑重要的基因和途径。通过各种数据集进行评估，儿童哮喘、特发性肺纤维化（IPF）和首发精神病(FEP)-PROMINENT在预测准确性和计算效率方面始终优于现有方法。突出还确定了涉及疾病机制的关键基因和途径。结论：PROMINENT代表了利用DNA甲基化数据进行表型预测的重大进步，在合理的计算时间内提供了高精度和可解释性。这种方法有望阐明复杂疾病的表观遗传基础，并增强DNA甲基化数据在生物医学研究中的应用。

{"title":"Pathway information on methylation analysis using deep neural network (PROMINENT): An interpretable deep learning method with pathway prior for phenotype prediction using gene-level DNA methylation","authors":"Soyeon Kim , Laizhi Zhang , Yidi Qin , Rebecca I. Caldino Bohn , Hyun Jung Park","doi":"10.1016/j.artmed.2025.103236","DOIUrl":"10.1016/j.artmed.2025.103236","url":null,"abstract":"<div><h3>Background</h3><div>DNA methylation is a key epigenetic marker that influences gene expression and phenotype regulation, and is affected by both genetic and environmental factors. Traditional linear regression methods such as elastic nets have been employed to assess the cumulative effects of multiple DNA methylation markers on phenotypes. However, these methods often fail to capture the complex nonlinear nature of the data. Recent deep learning approaches, such as MethylNet, have improved the prediction accuracy but lack interpretability and efficiency.</div></div><div><h3>Findings</h3><div>To address these limitations, we introduced <u>P</u>athway Info<u>r</u>mati<u>o</u>n on <u>M</u>ethylat<u>i</u>on Analysis using a Deep <u>Ne</u>ural <u>N</u>e<u>t</u>work (PROMINENT), a novel interpretable deep learning method that integrates gene-level DNA methylation data with biological pathway information for phenotype prediction. PROMINENT enhances interpretability and prediction accuracy by incorporating gene- and pathway-level priors from databases such as Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). It employs SHapley Additive exPlanations (SHAP) to prioritize significant genes and pathways. Evaluated across various datasets, childhood asthma, idiopathic pulmonary fibrosis (IPF), and first-episode psychosis (FEP)—PROMINENT consistently outperformed existing methods in terms of prediction accuracy and computational efficiency. PROMINENT also identified crucial genes and pathways involved in disease mechanisms.</div></div><div><h3>Conclusions</h3><div>PROMINENT represents a significant advancement in leveraging DNA methylation data for phenotype prediction, offering both high accuracy and interpretability within reasonable computational time. This method holds promise for elucidating the epigenetic underpinnings of complex diseases and enhancing the utility of DNA methylation data in biomedical research.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103236"},"PeriodicalIF":6.2,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145092830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

MMSupcon: An image fusion-based multi-modal supervised contrastive method for brain tumor diagnosis MMSupcon：一种基于图像融合的多模态监督对比脑肿瘤诊断方法

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-28 DOI: 10.1016/j.artmed.2025.103253

Haoyu Wang , Jing Zhang , Siying Wu , Haoran Wei , Xun Chen , Yunwei Ou , Xiaoyan Sun

The diagnosis of brain tumors is pivotal for effective treatment, with MRI serving as a commonly used non-invasive diagnostic modality in clinical practices. Fundamentally, brain tumor diagnosis is a type of pattern recognition task that requires the integration of information from multi-modal MRI images. However, existing fusion strategies are hindered by the scarcity of multi-modal imaging samples. In this paper, we propose a new training paradigm tailored for the scenario of multi-modal imaging in brain tumor diagnosis, called multi-modal supervised contrastive learning method (MMSupcon). This method significantly enhances diagnostic accuracy through two key components: multi-modal medical image fusion and multi-modal supervised contrastive loss. First, the fusion component integrates complementary imaging modalities to generate information-rich samples. Second, by introducing fused samples to guide original samples in learning feature consistency or inconsistency among classes, our loss component effectively preserves the integrity of cross-modal information while maintaining the distinctiveness of individual modalities. Finally, MMSupcon is validated on a real-world brain tumor dataset collected from Beijing Tiantan Hospital, achieving state-of-the-art performance. Furthermore, additional experiments on two public BraTS glioma classification datasets also demonstrate our substantial performance improvements. The source code is released at https://github.com/hywang02/MMSupcon.

脑肿瘤的诊断是有效治疗的关键，MRI在临床实践中是一种常用的无创诊断方法。从根本上说，脑肿瘤诊断是一种模式识别任务，需要整合来自多模态MRI图像的信息。然而，现有的融合策略受到多模态成像样本稀缺的阻碍。在本文中，我们提出了一种针对脑肿瘤诊断中的多模态成像场景量身定制的新的训练范式，称为多模态监督对比学习方法（MMSupcon）。该方法通过多模态医学图像融合和多模态监督对比损失两个关键部分显著提高了诊断准确率。首先，融合组件集成互补成像模式以生成信息丰富的样本。其次，通过引入融合样本来指导原始样本学习类之间的特征一致性或不一致性，我们的损失分量有效地保留了跨模态信息的完整性，同时保持了个体模态的独特性。最后，MMSupcon在北京天坛医院的真实脑肿瘤数据集上进行了验证，达到了最先进的性能。此外，在两个公开的BraTS胶质瘤分类数据集上进行的额外实验也证明了我们的性能有了实质性的提高。源代码发布在https://github.com/hywang02/MMSupcon。

{"title":"MMSupcon: An image fusion-based multi-modal supervised contrastive method for brain tumor diagnosis","authors":"Haoyu Wang , Jing Zhang , Siying Wu , Haoran Wei , Xun Chen , Yunwei Ou , Xiaoyan Sun","doi":"10.1016/j.artmed.2025.103253","DOIUrl":"10.1016/j.artmed.2025.103253","url":null,"abstract":"<div><div>The diagnosis of brain tumors is pivotal for effective treatment, with MRI serving as a commonly used non-invasive diagnostic modality in clinical practices. Fundamentally, brain tumor diagnosis is a type of pattern recognition task that requires the integration of information from multi-modal MRI images. However, existing fusion strategies are hindered by the scarcity of multi-modal imaging samples. In this paper, we propose a new training paradigm tailored for the scenario of multi-modal imaging in brain tumor diagnosis, called multi-modal supervised contrastive learning method (MMSupcon). This method significantly enhances diagnostic accuracy through two key components: multi-modal medical image fusion and multi-modal supervised contrastive loss. First, the fusion component integrates complementary imaging modalities to generate information-rich samples. Second, by introducing fused samples to guide original samples in learning feature consistency or inconsistency among classes, our loss component effectively preserves the integrity of cross-modal information while maintaining the distinctiveness of individual modalities. Finally, MMSupcon is validated on a real-world brain tumor dataset collected from Beijing Tiantan Hospital, achieving state-of-the-art performance. Furthermore, additional experiments on two public BraTS glioma classification datasets also demonstrate our substantial performance improvements. The source code is released at <span><span>https://github.com/hywang02/MMSupcon</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"170 ","pages":"Article 103253"},"PeriodicalIF":6.2,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145005191","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Privacy-preserving federated transfer learning for enhanced liver lesion segmentation in PET–CT imaging 保护隐私的联合迁移学习在PET-CT图像中增强肝脏病变分割

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-28 DOI: 10.1016/j.artmed.2025.103245

Rajesh Kumar , Shaoning Zeng , Jay Kumar , Zakria , Xinfeng Mao

Positron Emission Tomography-Computed Tomography (PET–CT) evolution is critical for liver lesion diagnosis. However, data scarcity, privacy concerns, and cross-institutional imaging heterogeneity impede accurate deep learning model deployment. We propose a Federated Transfer Learning (FTL) framework that integrates federated learning’s privacy-preserving collaboration with transfer learning’s pre-trained model adaptation, enhancing liver lesion segmentation in PET–CT imaging. By leveraging a Feature Co-learning Block (FCB) and privacy-enhancing technologies (DP, HE), our approach ensures robust segmentation without sharing sensitive patient data. (1) A privacy-preserving FTL framework combining federated learning and adaptive transfer learning; (2) A multi-modal FCB for improved PET–CT feature integration; (3) Extensive evaluation across diverse institutions with privacy-enhancing technologies like Differential Privacy (DP) and Homomorphic Encryption (HE). Experiments on simulated multi-institutional PET–CT datasets demonstrate superior performance compared to baselines, with robust privacy guarantees. The FTL framework reduces data requirements and enhances generalizability, advancing liver lesion diagnostics.

正电子发射断层扫描-计算机断层扫描（PET-CT）的演变对肝脏病变的诊断至关重要。然而，数据稀缺、隐私问题和跨机构成像异质性阻碍了准确的深度学习模型部署。我们提出了一个联邦迁移学习（FTL）框架，该框架将联邦学习的隐私保护协作与迁移学习的预训练模型适应相结合，增强了PET-CT成像中的肝脏病变分割。通过利用特征共同学习块（FCB）和隐私增强技术（DP， HE），我们的方法确保在不共享敏感患者数据的情况下进行稳健的分割。(1)结合联邦学习和自适应迁移学习的隐私保护FTL框架；(2)改进PET-CT特征集成的多模态FCB；(3)利用差分隐私（DP）和同态加密（HE）等隐私增强技术对不同机构进行广泛评估。在模拟的多机构PET-CT数据集上进行的实验表明，与基线相比，该方法具有优越的性能，并具有鲁棒的隐私保证。FTL框架减少了数据需求，增强了通用性，推进了肝脏病变诊断。

{"title":"Privacy-preserving federated transfer learning for enhanced liver lesion segmentation in PET–CT imaging","authors":"Rajesh Kumar , Shaoning Zeng , Jay Kumar , Zakria , Xinfeng Mao","doi":"10.1016/j.artmed.2025.103245","DOIUrl":"10.1016/j.artmed.2025.103245","url":null,"abstract":"<div><div>Positron Emission Tomography-Computed Tomography (PET–CT) evolution is critical for liver lesion diagnosis. However, data scarcity, privacy concerns, and cross-institutional imaging heterogeneity impede accurate deep learning model deployment. We propose a Federated Transfer Learning (FTL) framework that integrates federated learning’s privacy-preserving collaboration with transfer learning’s pre-trained model adaptation, enhancing liver lesion segmentation in PET–CT imaging. By leveraging a Feature Co-learning Block (FCB) and privacy-enhancing technologies (DP, HE), our approach ensures robust segmentation without sharing sensitive patient data. (1) A privacy-preserving FTL framework combining federated learning and adaptive transfer learning; (2) A multi-modal FCB for improved PET–CT feature integration; (3) Extensive evaluation across diverse institutions with privacy-enhancing technologies like Differential Privacy (DP) and Homomorphic Encryption (HE). Experiments on simulated multi-institutional PET–CT datasets demonstrate superior performance compared to baselines, with robust privacy guarantees. The FTL framework reduces data requirements and enhances generalizability, advancing liver lesion diagnostics.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"169 ","pages":"Article 103245"},"PeriodicalIF":6.2,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144922204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physical foundations for trustworthy medical imaging: A survey for artificial intelligence researchers 可信医学成像的物理基础：对人工智能研究人员的调查

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-26 DOI: 10.1016/j.artmed.2025.103251

Miriam Cobo , David Corral Fontecha , Wilson Silva , Lara Lloret Iglesias

Artificial intelligence in medical imaging has grown rapidly in the past decade, driven by advances in deep learning and widespread access to computing resources. Applications cover diverse imaging modalities, including those based on electromagnetic radiation (e.g., X-rays), subatomic particles (e.g., nuclear imaging), and acoustic waves (ultrasound). Each modality features and limitations are defined by its underlying physics. However, many artificial intelligence practitioners lack a solid understanding of the physical principles involved in medical image acquisition. This gap hinders leveraging the full potential of deep learning, as incorporating physics knowledge into artificial intelligence systems promotes trustworthiness, especially in limited data scenarios. This work reviews the fundamental physical concepts behind medical imaging and examines their influence on recent developments in artificial intelligence, particularly, generative models and reconstruction algorithms. Finally, we describe physics-informed machine learning approaches to improve feature learning in medical imaging.

在过去十年中，由于深度学习的进步和计算资源的广泛使用，医学成像领域的人工智能发展迅速。应用涵盖多种成像模式，包括基于电磁辐射（例如x射线）、亚原子粒子（例如核成像）和声波（超声波）的成像模式。每个模态的特征和限制都由其底层物理定义。然而，许多人工智能从业者对医学图像采集所涉及的物理原理缺乏扎实的理解。这一差距阻碍了充分利用深度学习的潜力，因为将物理知识纳入人工智能系统可以提高可信度，特别是在有限的数据场景中。这项工作回顾了医学成像背后的基本物理概念，并检查了它们对人工智能，特别是生成模型和重建算法的最新发展的影响。最后，我们描述了基于物理的机器学习方法，以改善医学成像中的特征学习。

{"title":"Physical foundations for trustworthy medical imaging: A survey for artificial intelligence researchers","authors":"Miriam Cobo , David Corral Fontecha , Wilson Silva , Lara Lloret Iglesias","doi":"10.1016/j.artmed.2025.103251","DOIUrl":"10.1016/j.artmed.2025.103251","url":null,"abstract":"<div><div>Artificial intelligence in medical imaging has grown rapidly in the past decade, driven by advances in deep learning and widespread access to computing resources. Applications cover diverse imaging modalities, including those based on electromagnetic radiation (e.g., X-rays), subatomic particles (e.g., nuclear imaging), and acoustic waves (ultrasound). Each modality features and limitations are defined by its underlying physics. However, many artificial intelligence practitioners lack a solid understanding of the physical principles involved in medical image acquisition. This gap hinders leveraging the full potential of deep learning, as incorporating physics knowledge into artificial intelligence systems promotes trustworthiness, especially in limited data scenarios. This work reviews the fundamental physical concepts behind medical imaging and examines their influence on recent developments in artificial intelligence, particularly, generative models and reconstruction algorithms. Finally, we describe physics-informed machine learning approaches to improve feature learning in medical imaging.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"169 ","pages":"Article 103251"},"PeriodicalIF":6.2,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144917649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

TIPs: Tooth instance and pulp segmentation based on hierarchical extraction and fusion of anatomical priors from cone-beam CT TIPs：基于锥形束CT解剖先验信息的分层提取和融合的牙实体和牙髓分割

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-23 DOI: 10.1016/j.artmed.2025.103247

Tao Zhong , Yang Ning , Xueyang Wu , Li Ye , Chichi Li , Yu Zhang , Yu Du

Accurate instance segmentation of tooth and pulp from cone-beam computed tomography (CBCT) images is essential but highly challenging due to the pulp’s small structures and indistinct boundaries. To address these critical challenges, we propose TIPs designed for Tooth Instance and Pulp segmentation. TIPs initially employs a backbone model to segment a binary mask of the tooth from CBCT images, which is then utilized to derive position prior of the tooth and shape prior of the pulp. Subsequently, we propose the Hierarchical Fusion Mamba models to leverage the strengths of both anatomical priors and CBCT images by extracting and integrating shallow and deep features from Convolution Neural Networks (CNNs) and State Space Sequence Models (SSMs), respectively. This process achieves tooth instance and pulp segmentation, which are then combined to obtain the final pulp instance segmentation. Extensive experiments on CBCT scans from 147 patients demonstrate that TIPs significantly outperforms state-of-the-art methods in terms of segmentation accuracy. Furthermore, we have encapsulated this framework into an openly accessible tool for one-click using. To our knowledge, this is the first toolbox capable of segmentation of tooth and pulp instances, with its performance validated on two external datasets comprising 59 samples from the Toothfairy2 dataset and 48 samples from the STS dataset. These results demonstrate the potential of TIPs as a practical tool to boost clinical workflows in digital dentistry, enhancing the precision and efficiency of dental diagnostics and treatment planning.

由于牙髓结构小，边界模糊，因此对牙髓和牙髓进行精确的分割是非常必要的，但也是非常具有挑战性的。为了解决这些关键的挑战，我们提出了为牙齿实例和牙髓分割设计的TIPs。TIPs首先使用骨干模型从CBCT图像中分割牙齿的二值掩模，然后利用该掩模获得牙齿的位置先验和牙髓的形状先验。随后，我们提出了分层融合曼巴模型，利用解剖先验和CBCT图像的优势，分别从卷积神经网络（cnn）和状态空间序列模型（SSMs）中提取和整合浅层和深层特征。该过程实现了牙体和牙髓的分割，然后将两者结合起来得到最终的牙髓分割。对147例患者的CBCT扫描进行的大量实验表明，TIPs在分割精度方面明显优于最先进的方法。此外，我们已经将这个框架封装成一个公开访问的工具，供一键使用。据我们所知，这是第一个能够分割牙齿和牙髓实例的工具箱，其性能在两个外部数据集上进行了验证，这些数据集包括来自Toothfairy2数据集的59个样本和来自STS数据集的48个样本。这些结果证明了TIPs作为一种实用工具的潜力，可以促进数字牙科临床工作流程，提高牙科诊断和治疗计划的准确性和效率。

{"title":"TIPs: Tooth instance and pulp segmentation based on hierarchical extraction and fusion of anatomical priors from cone-beam CT","authors":"Tao Zhong , Yang Ning , Xueyang Wu , Li Ye , Chichi Li , Yu Zhang , Yu Du","doi":"10.1016/j.artmed.2025.103247","DOIUrl":"10.1016/j.artmed.2025.103247","url":null,"abstract":"<div><div>Accurate instance segmentation of tooth and pulp from cone-beam computed tomography (CBCT) images is essential but highly challenging due to the pulp’s small structures and indistinct boundaries. To address these critical challenges, we propose TIPs designed for <u>T</u>ooth <u>I</u>nstance and <u>P</u>ulp <u>s</u>egmentation. TIPs initially employs a backbone model to segment a binary mask of the tooth from CBCT images, which is then utilized to derive position prior of the tooth and shape prior of the pulp. Subsequently, we propose the Hierarchical Fusion Mamba models to leverage the strengths of both anatomical priors and CBCT images by extracting and integrating shallow and deep features from Convolution Neural Networks (CNNs) and State Space Sequence Models (SSMs), respectively. This process achieves tooth instance and pulp segmentation, which are then combined to obtain the final pulp instance segmentation. Extensive experiments on CBCT scans from 147 patients demonstrate that TIPs significantly outperforms state-of-the-art methods in terms of segmentation accuracy. Furthermore, we have encapsulated this framework into an openly accessible tool for one-click using. To our knowledge, this is the first toolbox capable of segmentation of tooth and pulp instances, with its performance validated on two external datasets comprising 59 samples from the Toothfairy2 dataset and 48 samples from the STS dataset. These results demonstrate the potential of TIPs as a practical tool to boost clinical workflows in digital dentistry, enhancing the precision and efficiency of dental diagnostics and treatment planning.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"169 ","pages":"Article 103247"},"PeriodicalIF":6.2,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144917650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multiplex aggregation combining sample reweight composite network for pathology image segmentation 基于多重聚合的样本重权复合网络病理图像分割

IF 6.2 2区医学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Artificial Intelligence in Medicine

Pub Date : 2025-08-22 DOI: 10.1016/j.artmed.2025.103239

Dawei Fan , Zhuo Chen , Yifan Gao , Jiaming Yu , Kaibin Li , Yi Wei , Yanping Chen , Riqing Chen , Lifang Wei

In digital pathology, nuclei segmentation is a critical task for pathological image analysis, holding significant importance for diagnosis and research. However, challenges such as blurred boundaries between nuclei and background regions, domain shifts between pathological images, and uneven distribution of nuclei pose significant obstacles to segmentation tasks. To address these issues, we propose an innovative Causal inference inspired Diversified aggregation convolution Network named CDNet, which integrates a Diversified Aggregation Convolution (DAC), a Causal Inference Module (CIM) based on causal discovery principles, and a comprehensive loss function. DAC improves the issue of unclear boundaries between nuclei and background regions, and CIM enhances the model’s cross-domain generalization ability. A novel Stable-Weighted Combined loss function was designed that combined the chunk-computed Dice Loss with the Focal Loss and the Causal Inference Loss to address the issue of uneven nuclei distribution. Experimental evaluations on the MoNuSeg, GLySAC, and MoNuSAC datasets demonstrate that CDNet significantly outperforms other models and exhibits strong generalization capabilities. Specifically, CDNet outperforms the second-best model by 0.79% (mIoU) and 1.32% (DSC) on the MoNuSeg dataset, by 2.65% (mIoU) and 2.13% (DSC) on the GLySAC dataset, and by 1.54% (mIoU) and 1.10% (DSC) on the MoNuSAC dataset. Code is publicly available at https://github.com/7FFDW/CDNet.

在数字病理中，细胞核分割是病理图像分析的一项关键任务，对诊断和研究具有重要意义。然而，核与背景区域之间的边界模糊、病理图像之间的域转移以及核分布不均匀等挑战对分割任务构成了重大障碍。为了解决这些问题，我们提出了一种创新的基于因果推理的多元聚合卷积网络CDNet，该网络集成了多元聚合卷积（DAC）、基于因果发现原则的因果推理模块（CIM）和综合损失函数。DAC改进了核与背景区域边界不清的问题，CIM增强了模型的跨域泛化能力。为了解决核分布不均匀的问题，设计了一种新的稳定加权组合损失函数，将块计算的骰子损失与焦点损失和因果推理损失相结合。在MoNuSeg、GLySAC和MoNuSAC数据集上的实验评估表明，CDNet显著优于其他模型，并表现出强大的泛化能力。具体来说，CDNet在MoNuSeg数据集上比第二好的模型高出0.79% （mIoU）和1.32% (DSC)，在GLySAC数据集上高出2.65% （mIoU）和2.13% (DSC)，在MoNuSAC数据集上高出1.54% （mIoU）和1.10% （DSC）。代码可在https://github.com/7FFDW/CDNet上公开获取。

{"title":"Multiplex aggregation combining sample reweight composite network for pathology image segmentation","authors":"Dawei Fan , Zhuo Chen , Yifan Gao , Jiaming Yu , Kaibin Li , Yi Wei , Yanping Chen , Riqing Chen , Lifang Wei","doi":"10.1016/j.artmed.2025.103239","DOIUrl":"10.1016/j.artmed.2025.103239","url":null,"abstract":"<div><div>In digital pathology, nuclei segmentation is a critical task for pathological image analysis, holding significant importance for diagnosis and research. However, challenges such as blurred boundaries between nuclei and background regions, domain shifts between pathological images, and uneven distribution of nuclei pose significant obstacles to segmentation tasks. To address these issues, we propose an innovative Causal inference inspired Diversified aggregation convolution Network named CDNet, which integrates a Diversified Aggregation Convolution (DAC), a Causal Inference Module (CIM) based on causal discovery principles, and a comprehensive loss function. DAC improves the issue of unclear boundaries between nuclei and background regions, and CIM enhances the model’s cross-domain generalization ability. A novel Stable-Weighted Combined loss function was designed that combined the chunk-computed Dice Loss with the Focal Loss and the Causal Inference Loss to address the issue of uneven nuclei distribution. Experimental evaluations on the MoNuSeg, GLySAC, and MoNuSAC datasets demonstrate that CDNet significantly outperforms other models and exhibits strong generalization capabilities. Specifically, CDNet outperforms the second-best model by 0.79% (mIoU) and 1.32% (DSC) on the MoNuSeg dataset, by 2.65% (mIoU) and 2.13% (DSC) on the GLySAC dataset, and by 1.54% (mIoU) and 1.10% (DSC) on the MoNuSAC dataset. Code is publicly available at <span><span>https://github.com/7FFDW/CDNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"169 ","pages":"Article 103239"},"PeriodicalIF":6.2,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144895439","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0