Pub Date : 2026-01-31DOI: 10.1038/s41746-026-02393-z
Guoqiang Ren, Zhen Chen, Pengxiang Su, Da Li, Xiaoping Yang, Di Gai, Xin Wei, Weifeng Xu, Hongping Chen, Xiaoguang Zhao, Xiaofei Wang, Pengfei Liu, Honghua Ye, Yanfeng Ma
Accurate medical image segmentation continues to pose significant challenges, as existing methods often struggle to concurrently achieve efficient global context modeling, precise boundary delineation, and robust generalization. To address these issues, a novel framework named Contextual and Frequency-Guided Mamba Network (CFG-MambaNet) is presented. Specifically, a variable-scale state space block based on Mamba is employed so that long-range dependencies can be captured with linear complexity, efficiently addressing the inefficiency of Transformer-based models in high-resolution medical imaging. Moreover, a frequency-guided representation module is incorporated to explicitly separate global low-frequency structures from high-frequency boundary details, which significantly alleviates the difficulty of segmenting lesions with blurred contours or weak textures. Furthermore, an adaptive context aggregation mechanism is introduced to integrate heterogeneous semantic cues and to consistently highlight clinically critical regions, substantially improving robustness across diverse anatomical scales and morphologies. To further stabilize training and improve boundary adherence, a composite loss combined with deep supervision is employed. Extensive experiments were conducted on four publicly available datasets, including ACDC, Kvasir-SEG, ISIC, and SEED, covering cardiac MRI, endoscopy, dermoscopy, and pathology images.
{"title":"CFG-MambaNet: Contextual and Frequency-Guided Mamba Network for medical image segmentation","authors":"Guoqiang Ren, Zhen Chen, Pengxiang Su, Da Li, Xiaoping Yang, Di Gai, Xin Wei, Weifeng Xu, Hongping Chen, Xiaoguang Zhao, Xiaofei Wang, Pengfei Liu, Honghua Ye, Yanfeng Ma","doi":"10.1038/s41746-026-02393-z","DOIUrl":"https://doi.org/10.1038/s41746-026-02393-z","url":null,"abstract":"Accurate medical image segmentation continues to pose significant challenges, as existing methods often struggle to concurrently achieve efficient global context modeling, precise boundary delineation, and robust generalization. To address these issues, a novel framework named Contextual and Frequency-Guided Mamba Network (CFG-MambaNet) is presented. Specifically, a variable-scale state space block based on Mamba is employed so that long-range dependencies can be captured with linear complexity, efficiently addressing the inefficiency of Transformer-based models in high-resolution medical imaging. Moreover, a frequency-guided representation module is incorporated to explicitly separate global low-frequency structures from high-frequency boundary details, which significantly alleviates the difficulty of segmenting lesions with blurred contours or weak textures. Furthermore, an adaptive context aggregation mechanism is introduced to integrate heterogeneous semantic cues and to consistently highlight clinically critical regions, substantially improving robustness across diverse anatomical scales and morphologies. To further stabilize training and improve boundary adherence, a composite loss combined with deep supervision is employed. Extensive experiments were conducted on four publicly available datasets, including ACDC, Kvasir-SEG, ISIC, and SEED, covering cardiac MRI, endoscopy, dermoscopy, and pathology images.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"114 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146089871","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 : 2026-01-31DOI: 10.1038/s41746-026-02395-x
Damià Valero-Bover, David Monterde, Gerard Carot-Sans, Emili Vela, Rubèn González-Colom, Josep Roca, Caridad Pontes, Xabier Michelena, Maria Mercedes Nogueras, Pilar Aparicio, Inmaculada Corrales, Teresa Biec, Isaac Cano, Jordi Piera-Jiménez
Multimorbidity, a major driver of healthcare demand and clinical complexity, is often addressed in a disease-centric manner and remains insufficiently understood in its population-level dynamics. Using data from a 10-year population-based cohort of 5.5 million adults in Catalonia, Spain, we quantified multimorbidity-associated clinical complexity using the Adjusted Morbidity Groups (AMG) index to predict progression from low/moderate ( < P80) to high/very high ( ≥ P80) complexity. Machine learning models identified predictive factors, while network analyses explored co-occurrence patterns among chronic conditions. During follow-up, 39.2% of the individuals who remained alive throughout the analysis period transitioned to high/very high complexity. Baseline AMG score was the strongest predictor of progression, surpassing models relying solely on individual diagnoses. The most prevalent conditions were nutritional and endocrine disorders, anxiety, and hypertension, with notable sequential links between mental and physical disorders. Findings emphasize the need for integrated, patient-centred care strategies and population-based prevention approaches to mitigate multimorbidity progression.
{"title":"Ten-year population-based assessment of multimorbidity burden progression in a regional cohort of 5.5 million adults","authors":"Damià Valero-Bover, David Monterde, Gerard Carot-Sans, Emili Vela, Rubèn González-Colom, Josep Roca, Caridad Pontes, Xabier Michelena, Maria Mercedes Nogueras, Pilar Aparicio, Inmaculada Corrales, Teresa Biec, Isaac Cano, Jordi Piera-Jiménez","doi":"10.1038/s41746-026-02395-x","DOIUrl":"https://doi.org/10.1038/s41746-026-02395-x","url":null,"abstract":"Multimorbidity, a major driver of healthcare demand and clinical complexity, is often addressed in a disease-centric manner and remains insufficiently understood in its population-level dynamics. Using data from a 10-year population-based cohort of 5.5 million adults in Catalonia, Spain, we quantified multimorbidity-associated clinical complexity using the Adjusted Morbidity Groups (AMG) index to predict progression from low/moderate ( < P80) to high/very high ( ≥ P80) complexity. Machine learning models identified predictive factors, while network analyses explored co-occurrence patterns among chronic conditions. During follow-up, 39.2% of the individuals who remained alive throughout the analysis period transitioned to high/very high complexity. Baseline AMG score was the strongest predictor of progression, surpassing models relying solely on individual diagnoses. The most prevalent conditions were nutritional and endocrine disorders, anxiety, and hypertension, with notable sequential links between mental and physical disorders. Findings emphasize the need for integrated, patient-centred care strategies and population-based prevention approaches to mitigate multimorbidity progression.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"16 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146089870","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 : 2026-01-30DOI: 10.1038/s41746-026-02348-4
Maria Marloth,Celia Deane-Drummond,Philipp Kellmeyer,Marc Erich Latoschik,Jennifer A Chandler,Gerben Meynen,Kai Vogeley
Extended reality (=XR) provides promising opportunities for psychiatry in the future. However, psychiatric patients appear to be particularly vulnerable to virtual exposure. With a focus on virtual embodiment and virtual social interaction, we therefore, 1. describe the specific risks of virtual exposures, 2. discuss them in relation to specific psychopathological symptoms and 3. outline initial strategies that enable safe exposure with a strong emphasis on participatory designs.
{"title":"Anticipation and prevention of real risks of virtual environments in psychiatry.","authors":"Maria Marloth,Celia Deane-Drummond,Philipp Kellmeyer,Marc Erich Latoschik,Jennifer A Chandler,Gerben Meynen,Kai Vogeley","doi":"10.1038/s41746-026-02348-4","DOIUrl":"https://doi.org/10.1038/s41746-026-02348-4","url":null,"abstract":"Extended reality (=XR) provides promising opportunities for psychiatry in the future. However, psychiatric patients appear to be particularly vulnerable to virtual exposure. With a focus on virtual embodiment and virtual social interaction, we therefore, 1. describe the specific risks of virtual exposures, 2. discuss them in relation to specific psychopathological symptoms and 3. outline initial strategies that enable safe exposure with a strong emphasis on participatory designs.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"40 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146089060","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}
Endoplasmic reticulum stress-related cancer-associated fibroblasts (ERS-CAF) remodel the tumor microenvironment and drive immune exclusion and therapy resistance in chordoma, yet routine and non-invasive readouts of this biology are lacking. We hypothesized that standard pre-operative MRI and H&E whole-slide images (WSI) encode image-based surrogates of ERS-CAF-driven immunoregulation that can be learned and generalized across cancers. Three bulk-transcriptomic reference scores were defined for surrogate supervision, capturing ERS-program activity, ERS-CAF-immuneligand-receptor crosstalk and microenvironmental heterogeneity. In 126 chordoma cases, a stage-wise multimodal framework integrating calibrated WSI attention, gated radiopathomic fusion and domain alignment showed strong concordance with molecular profiles, independent prognostic value and biologically specific localization to fibrotic immune-excluded regions. These associations were generalized in zero-shot analyses to the TCGA pan-cancer. An MRI-only distilled model preserved most predictive performance with substantial gains in efficiency, supporting scalable non-invasive clinical application.
内质网应激相关的癌症相关成纤维细胞（ERS-CAF）重塑肿瘤微环境，驱动脊索瘤的免疫排斥和治疗抵抗，但缺乏这种生物学的常规和非侵入性解读。我们假设标准的术前MRI和H&E全片图像（WSI）编码了基于图像的ers - cafi驱动的免疫调节的替代物，可以在癌症中学习和推广。定义了三个大体积转录组参考评分，用于替代监督，捕获ers程序活性，ers - ca -免疫配体-受体串扰和微环境异质性。在126例脊索瘤病例中，一个分阶段的多模式框架整合了校准的WSI注意、门控的放射病理融合和区域比对，显示出与分子谱、独立的预后价值和纤维化免疫排斥区域的生物学特异性定位的强烈一致性。这些关联在零概率分析中被推广到TCGA泛癌。仅用mri提取的模型保留了大多数预测性能，并大幅提高了效率，支持可扩展的非侵入性临床应用。
{"title":"Decoding the ERS-CAF immunoregulatory axis via multimodal AI and its pan-cancer prognostic and therapeutic predictive value.","authors":"Bo-Wen Zheng,Chao Xia,Ming Tang,Wei Huang,Bo-Yv Zheng,Hua-Qing Niu,Jing Li,Tao-Lan Zhang,Hong Zhou,Ming-Xiang Zou","doi":"10.1038/s41746-026-02388-w","DOIUrl":"https://doi.org/10.1038/s41746-026-02388-w","url":null,"abstract":"Endoplasmic reticulum stress-related cancer-associated fibroblasts (ERS-CAF) remodel the tumor microenvironment and drive immune exclusion and therapy resistance in chordoma, yet routine and non-invasive readouts of this biology are lacking. We hypothesized that standard pre-operative MRI and H&E whole-slide images (WSI) encode image-based surrogates of ERS-CAF-driven immunoregulation that can be learned and generalized across cancers. Three bulk-transcriptomic reference scores were defined for surrogate supervision, capturing ERS-program activity, ERS-CAF-immuneligand-receptor crosstalk and microenvironmental heterogeneity. In 126 chordoma cases, a stage-wise multimodal framework integrating calibrated WSI attention, gated radiopathomic fusion and domain alignment showed strong concordance with molecular profiles, independent prognostic value and biologically specific localization to fibrotic immune-excluded regions. These associations were generalized in zero-shot analyses to the TCGA pan-cancer. An MRI-only distilled model preserved most predictive performance with substantial gains in efficiency, supporting scalable non-invasive clinical application.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"261 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146088913","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 : 2026-01-29DOI: 10.1038/s41746-026-02400-3
Yuqing Lu,Jing Chen,Nini Fan,Wenchao Song,Haiyang Sheng,Yinfeng Yang,Jinghui Wang
Drug-drug interaction (DDI) poses a major challenge in clinical pharmacology, often compromising therapeutic efficacy or causing serious adverse events. Traditional detection methods, heavily dependent on experimental assays and expert knowledge, are constrained by high costs and limited scalability. This work explores emerging machine learning (ML)-based strategies for predicting DDIs by leveraging the rapidly expanding biomedical data landscape. Recent advances in deep learning architectures, graph neural networks and sophisticated feature engineering have markedly improved predictive performance, offering scalable and data-efficient alternatives to conventional approaches. We further highlight real-world clinical applications where ML-based models have enhanced drug safety monitoring and informed therapeutic decision-making. Finally, we discuss critical challenges like model interpretability, generalizability and integration with clinical workflows, and outline future directions toward building robust, explainable and clinically actionable DDI prediction systems. This work provides a comprehensive perspective on how AI-driven methodologies are reshaping pharmacovigilance and precision therapeutics.
{"title":"Machine learning models for drug-drug interaction prediction from computational discovery to clinical application.","authors":"Yuqing Lu,Jing Chen,Nini Fan,Wenchao Song,Haiyang Sheng,Yinfeng Yang,Jinghui Wang","doi":"10.1038/s41746-026-02400-3","DOIUrl":"https://doi.org/10.1038/s41746-026-02400-3","url":null,"abstract":"Drug-drug interaction (DDI) poses a major challenge in clinical pharmacology, often compromising therapeutic efficacy or causing serious adverse events. Traditional detection methods, heavily dependent on experimental assays and expert knowledge, are constrained by high costs and limited scalability. This work explores emerging machine learning (ML)-based strategies for predicting DDIs by leveraging the rapidly expanding biomedical data landscape. Recent advances in deep learning architectures, graph neural networks and sophisticated feature engineering have markedly improved predictive performance, offering scalable and data-efficient alternatives to conventional approaches. We further highlight real-world clinical applications where ML-based models have enhanced drug safety monitoring and informed therapeutic decision-making. Finally, we discuss critical challenges like model interpretability, generalizability and integration with clinical workflows, and outline future directions toward building robust, explainable and clinically actionable DDI prediction systems. This work provides a comprehensive perspective on how AI-driven methodologies are reshaping pharmacovigilance and precision therapeutics.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"281 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072929","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}
To identify novel diagnostic biomarkers for acute pancreatitis (AP) and facilitate the early prediction of severe AP (SAP), this investigation characterized the serum metabolomic profiles of patients across distinct disease phases and integrated metabolomics with artificial intelligence to construct bile acid-based predictive models. The observational protocol was registered with the Chinese Clinical Trial Registry (ChiCTR2000034117) on June 24, 2020. Comparative metabolomic analysis revealed significant alterations in 303 metabolites and 461 lipid species in AP. Subsequent weighted gene coexpression network analysis demonstrated robust correlations between clinical parameters and specific metabolic clusters, particularly bile acids (BAs) and lipid species. Targeted quantification of 63 BAs was subsequently performed within a multicentre validation cohort (n = 948). Machine learning algorithms applied to these data facilitated the derivation of two distinct BA panels. The first panel, comprising nine BAs, demonstrated high diagnostic accuracy for AP, including among individuals with negative conventional enzymatic biomarkers, and effectively discriminated AP from acute cholangitis, as reflected by elevated area under the curve (AUC) values. A second panel, consisting of 13 BAs, reliably identified patients at elevated risk for SAP progression. Collectively, these results validate the translational potential of machine learning-driven metabolic biomarkers for the precision management of acute abdominal conditions, underscore the clinical utility of BAs as promising diagnostic and prognostic biomarkers in acute pancreatitis, and provide a new paradigm for the development of dynamic risk early-warning systems (Clinical Trial Registration Our study is an observational study registered in ChiCTR (ChiCTR2000034117) on 2020/06/24, not a prospective interventional clinical trial, and therefore does not fall under the ICMJE definition of a clinical trial requiring CONSORT compliance).
{"title":"Non-invasive liquid biopsy based on metabolomic profiling improves diagnosis and early warning of severe acute pancreatitis.","authors":"Dawei Deng,Qihang Yuan,Chen Pan,Junhong Chen,Jianjun Liu,Song Wei,Yi Liu,Yutong Zhu,Tianfu Wei,Jianliang Cao,Zeming Wu,Yuepeng Hu,Dong Shang,Peiyuan Yin","doi":"10.1038/s41746-025-02294-7","DOIUrl":"https://doi.org/10.1038/s41746-025-02294-7","url":null,"abstract":"To identify novel diagnostic biomarkers for acute pancreatitis (AP) and facilitate the early prediction of severe AP (SAP), this investigation characterized the serum metabolomic profiles of patients across distinct disease phases and integrated metabolomics with artificial intelligence to construct bile acid-based predictive models. The observational protocol was registered with the Chinese Clinical Trial Registry (ChiCTR2000034117) on June 24, 2020. Comparative metabolomic analysis revealed significant alterations in 303 metabolites and 461 lipid species in AP. Subsequent weighted gene coexpression network analysis demonstrated robust correlations between clinical parameters and specific metabolic clusters, particularly bile acids (BAs) and lipid species. Targeted quantification of 63 BAs was subsequently performed within a multicentre validation cohort (n = 948). Machine learning algorithms applied to these data facilitated the derivation of two distinct BA panels. The first panel, comprising nine BAs, demonstrated high diagnostic accuracy for AP, including among individuals with negative conventional enzymatic biomarkers, and effectively discriminated AP from acute cholangitis, as reflected by elevated area under the curve (AUC) values. A second panel, consisting of 13 BAs, reliably identified patients at elevated risk for SAP progression. Collectively, these results validate the translational potential of machine learning-driven metabolic biomarkers for the precision management of acute abdominal conditions, underscore the clinical utility of BAs as promising diagnostic and prognostic biomarkers in acute pancreatitis, and provide a new paradigm for the development of dynamic risk early-warning systems (Clinical Trial Registration Our study is an observational study registered in ChiCTR (ChiCTR2000034117) on 2020/06/24, not a prospective interventional clinical trial, and therefore does not fall under the ICMJE definition of a clinical trial requiring CONSORT compliance).","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"43 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146073002","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 : 2026-01-28DOI: 10.1038/s41746-026-02383-1
Júlia Tannús, Caroline Valentini, Eduardo Naves
Stroke is a leading cause of long-term disability, often affecting upper-limb motor function and requiring continuous assessment. The Fugl-Meyer Assessment (FMA), though a clinical gold standard, is time-consuming and demands specialized personnel. This study presents an AI-driven, low-cost rehabilitation exergame that simultaneously provides therapy and automatically estimates upper-limb motor performance during gameplay using only a standard camera. Sixteen kinematic and spatiotemporal features were extracted from 2D hand and arm trajectories of twelve post-stroke individuals (24 limbs, 14 affected) using the MediaPipe framework. Features such as hand angle, range of motion, movement area, traveled distance, and shoulder–elbow coordination showed strong correlations with FMA scores and stratified participants by motor severity. A lightweight linear regression model achieved high predictive performance (Spearman ρ = 0.92, R² = 0.89, RMSE = 4.42) and classified severity levels with 86–93% accuracy. This interpretable approach outperformed complex machine learning models, highlighting the clinical relevance of transparent metrics embedded in gameplay. The proposed framework is sensor-free, scalable, and reproducible, offering immediate feedback while reducing clinical workload and enabling accessible digital biomarkers for telerehabilitation and remote monitoring after stroke.
{"title":"AI-driven low-cost rehabilitation exergame as a lightweight framework for stroke assessment","authors":"Júlia Tannús, Caroline Valentini, Eduardo Naves","doi":"10.1038/s41746-026-02383-1","DOIUrl":"https://doi.org/10.1038/s41746-026-02383-1","url":null,"abstract":"Stroke is a leading cause of long-term disability, often affecting upper-limb motor function and requiring continuous assessment. The Fugl-Meyer Assessment (FMA), though a clinical gold standard, is time-consuming and demands specialized personnel. This study presents an AI-driven, low-cost rehabilitation exergame that simultaneously provides therapy and automatically estimates upper-limb motor performance during gameplay using only a standard camera. Sixteen kinematic and spatiotemporal features were extracted from 2D hand and arm trajectories of twelve post-stroke individuals (24 limbs, 14 affected) using the MediaPipe framework. Features such as hand angle, range of motion, movement area, traveled distance, and shoulder–elbow coordination showed strong correlations with FMA scores and stratified participants by motor severity. A lightweight linear regression model achieved high predictive performance (Spearman ρ = 0.92, R² = 0.89, RMSE = 4.42) and classified severity levels with 86–93% accuracy. This interpretable approach outperformed complex machine learning models, highlighting the clinical relevance of transparent metrics embedded in gameplay. The proposed framework is sensor-free, scalable, and reproducible, offering immediate feedback while reducing clinical workload and enabling accessible digital biomarkers for telerehabilitation and remote monitoring after stroke.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"52 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146057194","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}
Human–AI collaboration (H + AI) using large language models (LLMs) offers a promising approach to enhance clinical reasoning, documentation, and interpretation tasks. Following PRISMA 2020 (PROSPERO registration: CRD420251068272), we systematically compared H + AI with human-only (H) workflows, searching four databases through June 28, 2025. Ten peer-reviewed studies met eligibility criteria, with three preprints informing sensitivity analyses only. Diagnostic/interpretation accuracy (k = 2) showed a positive trend for H + AI (Risk Ratio [RR] 1.59), but was statistically imprecise and non-significant (95% CI 0.08 to 32.74), with 95% prediction intervals (PI) crossing the null. Composite diagnostic/management scores (k = 2) showed a statistically significant improvement (Mean Difference [MD] +4.88 percentage points, 95% CI + 0.65 to +9.12), yet the PI (–31.65 to 41.42) indicates high real-world uncertainty. Time efficiency (k = 3) showed no overall difference (MD + 0.4 min, 95%CI −4.18 to +4.97; I² = 70.1%). While documentation quality improved, but factual error rates remained high (~26–36%), undermining quality gains. In three-arm settings, H + AI did not universally outperform AI-only. Evidence remains preliminary yet highly uncertain and context-dependent. We recommend preregistered, pragmatic, multicenter trials embedded in real workflows, with harmonized core outcomes that prioritize safety/error metrics and interfaces that surface uncertainty and support verification.
{"title":"Human–large language model collaboration in clinical medicine: a systematic review and meta-analysis","authors":"Guoyong Wang, Kaijun Zhang, Jiyue Jiang, Chaonan Wang, Hui Bi, Haojun Liang, Zuoliang Qi, Ying Huang, Yu Li, Xiaonan Yang","doi":"10.1038/s41746-026-02382-2","DOIUrl":"https://doi.org/10.1038/s41746-026-02382-2","url":null,"abstract":"Human–AI collaboration (H + AI) using large language models (LLMs) offers a promising approach to enhance clinical reasoning, documentation, and interpretation tasks. Following PRISMA 2020 (PROSPERO registration: CRD420251068272), we systematically compared H + AI with human-only (H) workflows, searching four databases through June 28, 2025. Ten peer-reviewed studies met eligibility criteria, with three preprints informing sensitivity analyses only. Diagnostic/interpretation accuracy (k = 2) showed a positive trend for H + AI (Risk Ratio [RR] 1.59), but was statistically imprecise and non-significant (95% CI 0.08 to 32.74), with 95% prediction intervals (PI) crossing the null. Composite diagnostic/management scores (k = 2) showed a statistically significant improvement (Mean Difference [MD] +4.88 percentage points, 95% CI + 0.65 to +9.12), yet the PI (–31.65 to 41.42) indicates high real-world uncertainty. Time efficiency (k = 3) showed no overall difference (MD + 0.4 min, 95%CI −4.18 to +4.97; I² = 70.1%). While documentation quality improved, but factual error rates remained high (~26–36%), undermining quality gains. In three-arm settings, H + AI did not universally outperform AI-only. Evidence remains preliminary yet highly uncertain and context-dependent. We recommend preregistered, pragmatic, multicenter trials embedded in real workflows, with harmonized core outcomes that prioritize safety/error metrics and interfaces that surface uncertainty and support verification.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"183 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146057190","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}
Precise delineation of hepatocellular carcinoma (HCC) in abdominal CT is pivotal for early diagnosis and surgical planning, yet remains challenged by morphological heterogeneity, low contrast in small lesions, and scanner variability. To address these limitations, we propose Prompt-Mamba-AF, a framework tailored for robust HCC segmentation. Our method uniquely integrates anatomy-aware prompts to guide feature extraction within liver regions and leverages Mamba-based state-space modeling to capture long-range volumetric dependencies with linear complexity. Furthermore, we introduce structure-aware filtering to enforce topological consistency along lesion boundaries. Extensive validation on the LiTS, 3DIRCADb, and CHAOS benchmarks demonstrates that Prompt-Mamba-AF outperforms current state-of-the-art CNN and Transformer architectures. The model achieves leading Dice similarity and boundary accuracy while maintaining a compact parameter footprint (27.6M). Results indicate significant improvements in small nodule sensitivity and generalization across diverse imaging domains, positioning Prompt-Mamba-AF as an efficient solution for multi-center clinical workflows.
{"title":"Prompt-mamba filtering networks for accurate hepatocellular carcinoma lesion segmentation in abdominal CT.","authors":"Long Xia,Hai-Yang Chen,Ya-Wen Cao,Chen-Quan Gan,Jun-Zhang Zhao,Wei-Hua Zheng,Haiwen Jia,Shuai Jiang,Xuwang Li,Hua Li,Yi-Nuo Tu,Jun-Jing Zhang","doi":"10.1038/s41746-026-02371-5","DOIUrl":"https://doi.org/10.1038/s41746-026-02371-5","url":null,"abstract":"Precise delineation of hepatocellular carcinoma (HCC) in abdominal CT is pivotal for early diagnosis and surgical planning, yet remains challenged by morphological heterogeneity, low contrast in small lesions, and scanner variability. To address these limitations, we propose Prompt-Mamba-AF, a framework tailored for robust HCC segmentation. Our method uniquely integrates anatomy-aware prompts to guide feature extraction within liver regions and leverages Mamba-based state-space modeling to capture long-range volumetric dependencies with linear complexity. Furthermore, we introduce structure-aware filtering to enforce topological consistency along lesion boundaries. Extensive validation on the LiTS, 3DIRCADb, and CHAOS benchmarks demonstrates that Prompt-Mamba-AF outperforms current state-of-the-art CNN and Transformer architectures. The model achieves leading Dice similarity and boundary accuracy while maintaining a compact parameter footprint (27.6M). Results indicate significant improvements in small nodule sensitivity and generalization across diverse imaging domains, positioning Prompt-Mamba-AF as an efficient solution for multi-center clinical workflows.","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"117 1","pages":""},"PeriodicalIF":15.2,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146056415","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 : 2026-01-27DOI: 10.1038/s41746-026-02396-w
Florence D Mowlem, Jeremiah J Trudeau, Jill V Platko, Jos Bloemers, Emuella Flood, Jessica L Abel, Kelly Dumais, Paul O'Donohoe, Sabrina Grant, Ryan Naville-Cook, Onyekachukwu Illoh, Shannon D Keith, Jing Ju, Megan Fitter, Dorothee Oberdhan, Randall Winnette, Manuela Bossi, Naomi Suminski, Sonya Eremenco, Lindsay Hughes, Amy Fasiczka, Luisana Rojas, Michelle Campbell, Scottie Kern
{"title":"Publisher Correction: Best practice recommendations and considerations for designing and electronically implementing event-driven diaries in clinical trials.","authors":"Florence D Mowlem, Jeremiah J Trudeau, Jill V Platko, Jos Bloemers, Emuella Flood, Jessica L Abel, Kelly Dumais, Paul O'Donohoe, Sabrina Grant, Ryan Naville-Cook, Onyekachukwu Illoh, Shannon D Keith, Jing Ju, Megan Fitter, Dorothee Oberdhan, Randall Winnette, Manuela Bossi, Naomi Suminski, Sonya Eremenco, Lindsay Hughes, Amy Fasiczka, Luisana Rojas, Michelle Campbell, Scottie Kern","doi":"10.1038/s41746-026-02396-w","DOIUrl":"10.1038/s41746-026-02396-w","url":null,"abstract":"","PeriodicalId":19349,"journal":{"name":"NPJ Digital Medicine","volume":"9 1","pages":"77"},"PeriodicalIF":15.1,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12847838/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146065398","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：