Communications medicine最新文献

{"title":"Validation of remote multimodal AI screening for Parkinson disease across diverse settings.","authors":"Md Saiful Islam, Tariq Adnan, Abdelrahman Abdelkader, Zipei Liu, Evelyn Ma, Sooyong Park, Asif Azad, Pai Liu, Meghan Pawlik, Emily Hartman, Erin Shelton, Kristina B Larson, M Saifur Rahman, Cathe Schwartz, Karen Jaffe, Jamie L Adams, Ruth B Schneider, Jan Freyberg, E Ray Dorsey, Ehsan Hoque","doi":"10.1038/s43856-026-01606-6","DOIUrl":"https://doi.org/10.1038/s43856-026-01606-6","url":null,"abstract":"<p><strong>Background: </strong>Timely detection of Parkinson's disease (PD) remains limited by reliance on in-person neurological evaluations that are often costly and geographically inaccessible. To address these barriers, we develop PARK (Parkinson's Analysis with Remote Kinetic-tasks) - a web-based artificial intelligence (AI) tool that screens for PD using short webcam recordings of facial expression, motor, and speech tasks.</p><p><strong>Methods: </strong>Across eight independent studies (n = 1,865 participants; 670 with PD), participants completed three standardized tasks (smile mimicry, finger tapping, and pangram utterance) via webcam. Task-specific neural networks estimate PD risk and uncertainty, which are integrated through an uncertainty-calibrated fusion model (UFNet). Model performance is evaluated on one internal and two external test sets representing supervised and unsupervised real-world environments. Three movement disorder specialists also reviewed videos from 30 participants to benchmark clinical agreement of the PARK tool. User experience is assessed through structured surveys containing open-ended or multiple-choice questions.</p><p><strong>Results: </strong>PARK achieves accuracies of 80.2-80.6% and AUROC of 0.85-0.87 across all evaluation cohorts, with 83.3-86.5% sensitivity and 71.2-78.4% specificity. Predictive performance remains stable across sex, age, and ethnicity. Agreement with clinician judgments reaches Cohen's κ = 0.59. Uncertainty estimates reflect diagnostic confidence, and performance declines at high-uncertainty levels. Usability is rated highly (System Usability Scale  > 70) in both supervised and unsupervised settings, with low perceived risk and strong user preference for remote screening.</p><p><strong>Conclusions: </strong>PARK demonstrates promising accuracy and favorable user acceptance for remote PD screening, highlighting its potential as an accessible, equitable, and uncertainty-aware tool for neurological assessment when traditional care is challenging to obtain.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using an integrated omics approach to uncover the mechanisms underlying chemotherapy-induced peripheral neuropathy (CIPN).","authors":"Isabella Piga, Roberta Bonomo, Clizia Chinello, Lisa Pagani, Paola Pacifico, Cristina Meregalli","doi":"10.1038/s43856-026-01622-6","DOIUrl":"10.1038/s43856-026-01622-6","url":null,"abstract":"<p><p>Chemotherapy-induced peripheral neuropathy (CIPN) is a debilitating side effect with limited treatment options. The primary challenge in developing therapies is the lack of identified neurotoxic mechanisms. To address this, an integrated omics approach, combining transcriptomics, proteomics, and metabolomics, is essential to map the biological changes underlying the condition. Here, we show how data from these complementary approaches converge on key mechanisms involved in CIPN through a critical summary of animal and human studies. Current research highlights several key drivers of CIPN, such as inflammatory signaling and oxidative stress, mitochondrial dysfunction, and disrupted lipid metabolism. Although most data currently stem from preclinical models, the pathways identified offer promising targets for biomarker discovery and treatment. To translate these findings into clinical applications, integrated omics studies in human samples are urgently needed, focusing on a personalized approach. Future breakthroughs depend on large-scale human studies to tailor antineoplastic choices and neuroprotective treatments to individual patient needs.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"6 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13150028/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-scale data improves performance of machine learning model for long COVID identification.","authors":"Christopher Guardo, Zhang Xinmeng, Srushti Gangireddy, Yan Chao, V Eric Kerchberger, Alyson L Dickson, Emily R Pfaff, Hiral Master, Xin Yi, Melissa Basford, Christopher G Chute, Nguyen K Tran, Salvatore Mancuso, Toufeeq Ahmed Syed, Zhao Zhongming, Feng QiPing, Melissa Haendel, Christopher Lunt, Paul A Harris, Li Lang, Geoffrey S Ginsburg, Joshua C Denny, Dan M Roden, Wei Wei-Qi","doi":"10.1038/s43856-026-01621-7","DOIUrl":"https://doi.org/10.1038/s43856-026-01621-7","url":null,"abstract":"<p><strong>Background: </strong>Long COVID affects a substantial proportion of the over 778 million individuals infected with SARS-CoV-2, yet predictive models remain limited in scope. While existing efforts, such as the National COVID Cohort Collaborative (N3C), have leveraged electronic health record (EHR) data for risk prediction and identification, accumulating evidence points to additional contributions from social, behavioral, and genetic factors.</p><p><strong>Methods: </strong>Using a diverse cohort of SARS-CoV-2-infected individuals (n > 17,200) from the NIH All of Us Research Program, we investigated whether integrating EHR data with survey-based and genomic information improves model performance.</p><p><strong>Results: </strong>Our multi-scale approach outperforms EHR-only model's area under the receiver operating curve 0.736 (95% CI: 0.730, 0.741), achieving an area of 0.748 (0.741,0.755). Among the top predictors, active-duty service status, and self-reported fatigue are the most informative survey features.</p><p><strong>Conclusions: </strong>These findings highlight the importance of incorporating multi-scale data to improve risk stratification and inform personalized interventions for long COVID. However the relative increase in accuracy is modest, and the cost of collecting genetic and survey data should be considered before implementation.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147847000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Relationship between socioeconomic inequality and multimorbidity progression in UK Biobank data.","authors":"Jie Zhang, Lasse Bjerg, Susanne Boel Graversen, Henrik Støvring, Christina C Dahm, Luke Johnston, Bendix Carstensen, Daniel R Witte","doi":"10.1038/s43856-026-01607-5","DOIUrl":"https://doi.org/10.1038/s43856-026-01607-5","url":null,"abstract":"<p><strong>Background: </strong>Socio-economic status (SES) is associated with many adverse health outcomes, yet it remains unclear how SES relates to the rate at which people accumulate long-term conditions (LTCs) over time. We investigated this relationship between SES and disease accumulation using longitudinal disease tracking data.</p><p><strong>Methods: </strong>We analyzed data from the UK Biobank study (n = 502,368, median age 58 years [range 37-73], 46% male at baseline) with a median follow-up of 15.8 years. We tracked accumulation of 80 specified LTCs (identified from hospital records using ICD-10 codes). Multistate models were used to estimate the transition rates between SES and incremental morbidity states (i.e., 0 to 1 LTC, 1 to 2 LTCs, until 7 to 8 + LTCs), with death as the absorbing state. SES indicators included education level, family income, Townsend Deprivation Index, and Index of Multiple Deprivation. The models were adjusted for age, sex, ethnicity, calendar year, current number of LTCs, and lifestyle factors.</p><p><strong>Results: </strong>Over 7.5 million person-years of follow-up, we observe a clear socioeconomic gradient in disease accumulation rates. All four SES indicators are associated with accelerated morbidity progression and mortality. The socioeconomic gradient is evident across all transition stages but notably stronger for the initial transition from health to the first LTC, where the lowest income group has a 71% higher transition rate (95% CI: 1.67-1.76).</p><p><strong>Conclusion: </strong>Disadvantaged SES is associated with higher rates of progression to subsequent morbidities. These findings show the lasting impact of socioeconomic disadvantages on the widening health gap in later adulthood.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846950","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comprehensive evaluation of cross cancer generalization in histopathology segmentation models across 21 tumor types.","authors":"Tillmann Bedau, Christian Harder, Abdulkader Al-Shughri, Yuan Wang, Alexey Pryalukhin, Marie-Lisa Eich, Su Ir Lyu, Reinhard Büttner, Alexander Quaas, Yuri Tolkach","doi":"10.1038/s43856-026-01601-x","DOIUrl":"https://doi.org/10.1038/s43856-026-01601-x","url":null,"abstract":"<p><strong>Background: </strong>Artificial intelligence has significantly advanced computational pathology by enabling high-resolution, clinical-grade tumor segmentation models with state-of-the-art diagnostic accuracy. Creating such models is resource-intensive, requiring substantial time and domain expertise. Additionally, deep learning models are typically restricted to single tumor types, making it challenging to develop separate models for each tumor type. Cross-cancer generalization of segmentation models could address this bottleneck and pave the way for pan-cancer segmentation models.</p><p><strong>Methods: </strong>We evaluated the cross-tumor generalization capability of five tissue segmentation models (breast, colon, lung, kidney, prostate) using 21 cancer types from The Cancer Genome Atlas, totaling over 7,700 whole slide images. Representative large tumor and benign regions were manually selected, and segmentation accuracy was evaluated using a semiquantitative scale (0-10).</p><p><strong>Results: </strong>Here we show that the lung model demonstrates excellent cross-cancer performance (overall mean score 7.9 ± 2.1), effectively segmenting tumor regions in many non-lung cancers with segmentation accuracy similar to its native domain in 11 of 19 other epithelial tumors and melanoma, achieving particularly strong results in ovarian cancer (9.2 ± 0.9). The breast and colon models also show strong cross-domain performance, while the kidney and prostate models exhibit more limited generalization. Overall, high-precision segmentation is achievable in most cancer types using existing models.</p><p><strong>Conclusions: </strong>Existing segmentation models generalize across multiple cancer types, reducing the need to develop new, entity-specific models from scratch. This cross-domain generalization enables fast-track model development and supports future creation of robust pan-cancer segmentation models. Leveraging these capabilities could accelerate clinical integration of pathology artificial intelligence tools and enable reproducible biomarker discovery.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Implications for distancing measures on in-person school and work attendance from Italian post-pandemic social contact data.","authors":"Lorenzo Lucchini, Valentina Marziano, Filippo Trentini, Chiara Chiavenna, Elena D'Agnese, Vittoria Offeddu, Mattia Manica, Piero Poletti, Duilio Balsamo, Giorgio Guzzetta, Marco Ajelli, Alessia Melegaro, Stefano Merler","doi":"10.1038/s43856-026-01543-4","DOIUrl":"https://doi.org/10.1038/s43856-026-01543-4","url":null,"abstract":"<p><strong>Background: </strong>The collection of updated post-COVID-19 data on social contacts is critical for future epidemiological assessment and evaluation of non-pharmaceutical interventions.</p><p><strong>Methods: </strong>We conducted two waves of an online survey in Italy (March 2022 and March 2023), collecting representative data on direct (verbal/physical) and indirect (indoor co-location) contacts. Using a generalised linear mixed model, we analysed social contact determinants and the impact of work-from-home and distance learning on reducing a pathogen's reproduction number (R). Additionally, we calibrated an age-structured model to the 2023-2024 influenza A epidemic in Italy to explore the impact of alternative in-person attendance scenarios on infection attack rates.</p><p><strong>Results: </strong>We find that in-person attendance significantly increases contacts: adults attending in person have 1.69 times (95%CI: 1.55-1.83) more contacts than those staying home, while children/adolescents 2.36 (95%CI: 1.96-2.84). Limiting in-person work alone marginally affects R, whereas combining work-from-home with distance learning (from primary school onwards) reduces R by up to 23.2% (95%CI: 13.7-30.1%), with minimal additional benefit from suspending early childcare. In the influenza A case study, seasonal infection attack rates range from 14.7% (95%PI: 12.8-16.5%) under full in-person attendance to <0.2% under the most restrictive scenario. Moderate interventions (suspension of tertiary education and work-from-home) reduce attack rates by up to one fourth among adults (15-64 years) and one sixth among older individuals.</p><p><strong>Conclusions: </strong>This study provides post-pandemic contact matrices for Italy, essential for modelling transmission of respiratory pathogens, and quantitative evidence on the epidemiological impact of targeted physical distancing measures, thereby supporting future policy design.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147847046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Geometric deep learning enables high-fidelity network imputation for HIV transmission modeling.","authors":"Steven J Clipman, Jasmine Wang, Shruti H Mehta, Shobha Mohapatra, Aylur K Srikrishnan, Muniratnam Suresh Kumar, Gregory M Lucas, Carl A Latkin, Sunil S Solomon","doi":"10.1038/s43856-026-01640-4","DOIUrl":"https://doi.org/10.1038/s43856-026-01640-4","url":null,"abstract":"<p><strong>Background: </strong>Accurately mapping social and risk networks is critical for understanding and controlling infectious disease transmission, especially among hard-to-reach populations, such as people who inject drugs (PWID). Yet, empirical sociometric network ascertainment remains challenging and resource-intensive. Geometric deep learning may provide a scalable approach to inferring network structure from individual-level data, but its real-world performance and translation to epidemic modeling remain undercharacterized.</p><p><strong>Methods: </strong>We trained a graph neural network (GNN) to predict injection partnerships from a longitudinal network study of 2512 PWID in New Delhi, India, using demographic, behavioral, and spatial injection-venue features. We compared the GNN with exponential random graph models (ERGMs), evaluated structural similarity between empirical and imputed networks, and assessed validity in an independent PWID network. To examine translational utility, we calibrated a network-based HIV transmission model on either the empirical or GNN-imputed injection network and compared HIV incidence two years after scaling interventions across venues.</p><p><strong>Results: </strong>The GNN achieves balanced predictive performance (accuracy 60.8%, precision 59.4%, recall 67.9%, F1 63.4%), outperforming ERGMs, and yields an imputed network with structural concordance to the empirical network (spectral similarity 0.87). Incorporating venue data increases accuracy from 51.0% to 60.8%. In the external cohort, the GNN maintains performance (F1: 61.3%) and captures structural changes. In the HIV model, calibrating on the GNN-imputed versus empirical network produces incidence curves that differ by at most 0.4 infections per 100 person-years.</p><p><strong>Conclusions: </strong>GNN-based network imputation can recover sufficient epidemiologically relevant network structure to preserve conclusions about HIV interventions, illustrating how geometric deep learning can support network-informed epidemic modeling when full sociometric ascertainment is infeasible.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-driven identification of hub chains in disease accruals to reduce healthcare utilization and costs.","authors":"Shasha Han, Can Zhou, Sairan Li, Shuya Zhou, Ruitai Shao, Weizhong Yang, Chen Wang","doi":"10.1038/s43856-026-01617-3","DOIUrl":"https://doi.org/10.1038/s43856-026-01617-3","url":null,"abstract":"<p><strong>Background: </strong>Understanding how diseases accumulate over time is crucial for managing their combined impact. We aim to identify these common pathways of multimorbidity and determine how they increase healthcare use and costs beyond the sum of single diseases.</p><p><strong>Methods: </strong>We analyzed hospital records from over 600,000 patients in China between 2013 and 2021. We identified significant sequences of three diseases diagnosed over time and measured their synergistic effect by comparing the total healthcare use along a sequence to the added use of each disease occurring in isolation. Outcomes included hospital visits, length of stay, total costs, and out-of-pocket spending. We then grouped these sequences to identify common, high-impact connections between diseases, referred to as hub chains.</p><p><strong>Results: </strong>Here we show that distinct disease sequences are common, with 83 pathways identified in women and 174 in men. The combined cost of a disease sequence often exceeds the sum of its parts, in 69.9% (58/83) of female and 81.0% (141/174) of male trajectories, demonstrating a synergistic effect. Pathways involving cancer-related care show the highest resource use for both sexes. We find that a small number of critical hub chains connect many sequences; analyses indicate that interrupting these hubs could reduce associated healthcare use and costs by approximately half.</p><p><strong>Conclusions: </strong>This study demonstrates that specific sequences of diseases, particularly those linked by hub chains, drive disproportionate healthcare burdens. Focusing clinical interventions and policy on these high-impact pathways offers a promising strategy to alleviate system strain and improve care for patients with multimorbidity.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846959","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Acute SARS-CoV-2 infection and self-reported post-acute cognitive dysfunctions from the Danish EFTER-COVID survey.","authors":"Nete Munk Nielsen, Lampros Spiliopoulos, Anna Irene Vedel Sørensen, Elisabeth O'Regan, Peter Bager, Steen Ethelberg, Anders Koch, Poul Videbech, Anders Hviid","doi":"10.1038/s43856-025-01323-6","DOIUrl":"10.1038/s43856-025-01323-6","url":null,"abstract":"<p><strong>Background: </strong>The extent and burden of post-acute cognitive dysfunctions following SARS-CoV-2 infection is uncertain.</p><p><strong>Methods: </strong>25,485 SARS-CoV-2 test-positive and 25,032 test-negative individuals were repeatedly asked to score symptoms of subjective cognitive deficits 2 to 18 months after test using the \"Cognitive complaints in bipolar disorder rating assessment\" (COBRA) tool. Poisson mixed-effects models were used to estimate Score Ratios (SRs) by comparing scores between test-positive and test-negative individuals.</p><p><strong>Results: </strong>At each follow-up point, test-positive individuals have low but slightly higher mean COBRA scores compared with test-negatives. For the combined 2-18 months period, COBRA scores among test-positive individuals are 11% higher than corresponding scores among test-negatives (SR_2-18mth = 1.11 (95% CI; 1.09-1.13)). Of effect modifiers explored, being hospitalized with a positive SARS-CoV-2 test particularly elevates COBRA scores (SR_2-18mth = 1.38 (95% CI; 1.24-1.54)).</p><p><strong>Conclusion: </strong>In the general population of SARS-CoV-2 infected individuals, self-reported post-acute scores of cognitive dysfunctions are low and only slightly higher than corresponding scores among test-negatives. Higher COBRA scores among hospitalized SARS-CoV-2 test positives corroborate with long-term cognitive impairment being most pronounced among those with severe SARS-CoV-2 infection.</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":"6 1","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13136329/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147824268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Climate and socioeconomic factors drive heterogeneous dengue risk escalation in the Chinese population.","authors":"Xu Guang, Yifei He, Mengjie Geng, Meifang Liu, Jia Wan, Dongfeng Kong, Zhen Zhang, Lanbin Xiang, Liangqiang Lin, Rongxin He, Ning Zhang, Felipe Arley Costa Pessoa, Claudia Maria Ríos Velasquez, Carolina Mercedes Laurent Singh, Pritesh Lalwani, Jie Huang, Haidong Wang, Jue Liu, Bin Zhu","doi":"10.1038/s43856-026-01628-0","DOIUrl":"https://doi.org/10.1038/s43856-026-01628-0","url":null,"abstract":"<p><strong>Background: </strong>Dengue risk is increasingly shaped by climate change and rapid urbanization, yet comprehensive, multidimensional risk assessments grounded in a One Health perspective remain scare.</p><p><strong>Methods: </strong>We develop a geographically eXplainable artificial intelligence (GeoXAI) model to estimate dengue hazard across China in 2024 (current), 2050, and 2100 under different shared socioeconomic pathway (SSP) scenarios. A hazard-exposure-vulnerability framework is then used to assess dengue risk by integrating dengue hazard, human exposure, and social vulnerability.</p><p><strong>Results: </strong>Here we show a northward expansion of high-hazard areas, with the minimum temperature in the coldest month being the dominant driver (27.2% contribution). Moreover, socioeconomic factors such as population density (3.8%) and urbanization level (2.7%) will further amplify dengue hazard. Current dengue risk assessments reveal high-risk clusters in Southwest China and megacities. Dengue risk exhibits spatially heterogeneous escalation in the future, with Southwest and Southeast China facing the steepest growth and Northwest China experiencing disproportionate increases. Compared to the current, dengue risk in SSP585-a high greenhouse gas emission scenario and limited climate policy interventions-increases by 6.01% (2050) and 8.21% (2100), representing the largest escalation among the three SSPs.</p><p><strong>Conclusions: </strong>Despite ongoing disease control efforts, our findings underscore the need to intensify integrated surveillance and multidimensional intervention strategies against escalating dengue risk in China, and offers lessons for other prevalent Aedes-borne diseases (e.g., chikungunya).</p>","PeriodicalId":72646,"journal":{"name":"Communications medicine","volume":" ","pages":""},"PeriodicalIF":5.4,"publicationDate":"2026-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147824295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}