Nature Medicine最新文献

{"title":"Real-world deployment of a fine-tuned pathology foundation model for lung cancer biomarker detection","authors":"Gabriele Campanella, Neeraj Kumar, Swaraj Nanda, Siddharth Singi, Eugene Fluder, Ricky Kwan, Silke Muehlstedt, Nicole Pfarr, Peter J. Schüffler, Ida Häggström, Noora Neittaanmäki, Levent M. Akyürek, Alina Basnet, Tamara Jamaspishvili, Michel R. Nasr, Matthew M. Croken, Fred R. Hirsch, Arielle Elkrief, Helena Yu, Orly Ardon, Gregory M. Goldgof, Meera Hameed, Jane Houldsworth, Maria Arcila, Thomas J. Fuchs, Chad Vanderbilt","doi":"10.1038/s41591-025-03780-x","DOIUrl":"https://doi.org/10.1038/s41591-025-03780-x","url":null,"abstract":"<p>Artificial intelligence models using digital histopathology slides stained with hematoxylin and eosin offer promising, tissue-preserving diagnostic tools for patients with cancer. Despite their advantages, their clinical utility in real-world settings remains unproven. Assessing EGFR mutations in lung adenocarcinoma demands rapid, accurate and cost-effective tests that preserve tissue for genomic sequencing. PCR-based assays provide rapid results but with reduced accuracy compared with next-generation sequencing and require additional tissue. Computational biomarkers leveraging modern foundation models can address these limitations. Here we assembled a large international clinical dataset of digital lung adenocarcinoma slides (<i>N</i> = 8,461) to develop a computational EGFR biomarker. Our model fine-tunes an open-source foundation model, improving task-specific performance with out-of-center generalization and clinical-grade accuracy on primary and metastatic specimens (mean area under the curve: internal 0.847, external 0.870). To evaluate real-world clinical translation, we conducted a prospective silent trial of the biomarker on primary samples, achieving an area under the curve of 0.890. The artificial-intelligence-assisted workflow reduced the number of rapid molecular tests needed by up to 43% while maintaining the current clinical standard performance. Our retrospective and prospective analyses demonstrate the real-world clinical utility of a computational pathology biomarker.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"21 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144586599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Recombinant monoclonal antibody siltartoxatug versus plasma-derived human tetanus immunoglobulin for tetanus: a randomized, double-blind, active-controlled, phase 3 trial","authors":"Zijing Liang, Si Liu, Wei Guo, Zhe Deng, Wenkai Bin, Anyong Yu, Junyan Hu, Lidong Wu, Zhanfei Li, Wei Huang, He Li, Dapeng Cheng, Shugui Li, Qinghao Guo, Dongshan Zhang, Xinming Yan, Chunlei Wang, Wenwei Cai, Banghan Ding, Wenqiang Li, Xu Li, Bin Xu, Lei He, Yanhong Ouyang, Hong Zhan, Jianwei Wang, Yan Zhao, Xinyu Liu, Wenxi Xiang, Meizhuo Zhang, Zhihua Zhang, Jiyuan Ding, Xiaohu Kuang, Weihong Zheng, Huaxin Liao, Wanmei Wang, Chuanlin Wang","doi":"10.1038/s41591-025-03791-8","DOIUrl":"https://doi.org/10.1038/s41591-025-03791-8","url":null,"abstract":"<p>Tetanus remains an important global public health concern. Currently, the only recommended passive immunization therapy for tetanus prophylaxis is plasma-derived human tetanus immunoglobulin (HTIG), which faces a global supply shortage and can transmit infectious pathogens. Despite not being endorsed by WHO due to safety concerns, equine tetanus antitoxin remains widely used in some countries. We conducted a randomized, double-blind, phase 3 trial to evaluate siltartoxatug—a first-in-class recombinant monoclonal antibody—for tetanus postexposure prophylaxis. Participants (<i>n</i> = 675) were randomized (2:1) to receive a single intramuscular injection of siltartoxatug 10 mg or HTIG 250 IU. The study met its primary outcome, with siltartoxatug demonstrating superiority to HTIG in the proportion of participants with an increase of anti-tetanus neutralizing antibody titers from baseline (ΔTiter) ≥ 0.01 IU ml⁻¹ (95.4% versus 53.2%; intergroup difference 42.3% (95% confidence interval, 35.5–49.1; <i>P</i> &lt; 0.0001)). The safety profiles were comparable, with similar incidence of adverse events between the siltartoxatug (38.2%, 168 of 440) and HTIG (33.9%, 75 of 221) groups. These findings highlight siltartoxatug as an effective and safe option for passive immunization against tetanus. ClinicalTrials.gov registration: NCT05664750.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"109 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144577736","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global lifetime estimates of expected and preventable gastric cancers across 185 countries","authors":"Jin Young Park, Damien Georges, Catharina J. Alberts, Freddie Bray, Gary Clifford, Iacopo Baussano","doi":"10.1038/s41591-025-03793-6","DOIUrl":"https://doi.org/10.1038/s41591-025-03793-6","url":null,"abstract":"<p>Chronic infection with <i>Helicobacter pylori</i> is a modifiable cause of gastric cancer. To assist policymakers in advocating for and planning prevention strategies, we projected the future burden of gastric cancer, including that attributable to <i>H. pylori</i>, among a cohort of young people born in 2008–2017. Expected gastric cancer cases, in the absence of intervention, were quantified in 185 countries by combining national age-specific incidence rates from GLOBOCAN 2022 and cohort-specific mortality rates from the United Nations’ demographic projections. Globally, 15.6 million (95% uncertainty interval 14.0–17.3 million) lifetime gastric cancer cases are expected within these birth cohorts, 76% of which are attributable to <i>H. pylori</i>. Two-thirds of cases will be concentrated in Asia, followed by the Americas and Africa. Whereas 58% of cases are expected in traditionally high-incidence areas for gastric cancer, 42% of cases are expected to occur in lower-incidence areas owing to demographic changes, particularly in sub-Saharan Africa, where the future burden could be six times greater than estimated in 2022. A shift in focus toward the life course of today’s young people and their prospects of developing gastric cancer, with or without effective interventions, underscores the need for greater investment in gastric cancer prevention, including the implementation of population-based <i>H. pylori</i> screen-and-treat strategies.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"37 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144577817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thirty years of Europe’s centralized procedure for approving medicines","authors":"Emer Cooke, Bruno Sepodes","doi":"10.1038/s41591-025-03826-0","DOIUrl":"https://doi.org/10.1038/s41591-025-03826-0","url":null,"abstract":"<p>Few things epitomize the European Medicines Agency (EMA) as much as its now 30-year-old centralized procedure for approving human medicines. With one submission and one evaluation, a medicine can be authorized simultaneously in all 27 European Union (EU) member states.</p><p>As the EMA marks its 30th anniversary in 2025 with a year-long series of events at its offices in Amsterdam, it will be reflecting on the far-reaching achievements of the centralized procedure. But it is also a moment for renewed advocacy for international cooperation.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"37 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144568375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance of the American Heart Association's PREVENT risk score for cardiovascular risk prediction in a multiethnic population.","authors":"Roy O Mathew, Sadiya S Khan, Katherine R Tuttle, Jennifer E Ho, Dmitry Abramov, Sripal Bangalore, Mandeep S Sidhu, Chiadi E Ndumele, Tiffany M Powell-Wiley, Ian J Neeland, Josef Coresh, Mitchell S V Elkind, Janani Rangaswami","doi":"10.1038/s41591-025-03789-2","DOIUrl":"https://doi.org/10.1038/s41591-025-03789-2","url":null,"abstract":"<p><p>The Predicting Risk of Cardiovascular EVENTS (PREVENT) equations, created and endorsed by the American Heart Association, provide cardiovascular risk estimates for the general population, but have not yet been tested in multiethnic populations. In the present study, in a large nationwide multiethnic sample of US veterans, the utility of PREVENT to predict the risk of total cardiovascular disease (CVD: fatal and nonfatal myocardial infarction, stroke or heart failure; PREVENT-CVD), atherosclerotic cardiovascular disease (ASCVD: fatal and nonfatal myocardial infarction or stroke; PREVENT ASCVD) and heart failure was evaluated. We assessed the discrimination and calibration performance of ASCVD prediction with PREVENT ASCVD compared with a previous risk-prediction score, pooled cohort equations (PCEs). Among 2,500,291 veterans aged 30-79 years (93.9% men and 6.1% women), 407,342 total CVD events occurred over a median (interquartile range (IQR)) follow-up of 5.8 (IQR = 3.1-8.3) years. The Concordance index (C-index) (95% confidence interval (CI)) for PREVENT-CVD was 0.65 (95% CI = 0.65-0.65) in the overall sample and was similar across different race and ethnic groups (Asian, Native Hawaiian or Pacific Islander, 0.70 (95% CI = 0.69-0.71); Hispanic, 0.70 (95% CI = 0.69-0.70); non-Hispanic Black. 0.68 (95% CI = 0.68-0.69) and non-Hispanic White, 0.65 (95% CI = 0.64-0.65)). C-indices were similar between PREVENT ASCVD and PCEs and ranged from 0.61 to 0.63. Calibration slopes for PREVENT-CVD and -ASCVD in the overall sample were 1.09 (s.e. = 0.04) and 1.15 (s.e. = 0.04), respectively. In contrast, PCEs demonstrated overprediction for ASCVD with a calibration slope of 0.51 (s.e. = 0.06). Calibration slopes for PREVENT and PCEs were similar across race and ethnic groups. Among US veterans, the PREVENT equations accurately estimated CVD and ASCVD risk with some variability across race and ethnicity groups and outperformed PCEs for ASCVD risk prediction.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":" ","pages":""},"PeriodicalIF":58.7,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144564985","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From one to many gene therapies","authors":"","doi":"10.1038/d41591-025-00043-7","DOIUrl":"https://doi.org/10.1038/d41591-025-00043-7","url":null,"abstract":"Instead of making one-off treatments, industrialized gene therapy might be able to treat multiple diseases.","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"46 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144547491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Trials for LLM-supported clinical decisions in African primary healthcare","authors":"Bilal A. Mateen, Vaishnavi Menon, Ambrose Agweyu, Robert Korom, Elizabeth Omoluabi, David McAfee, Natnael Shimelash, Samuel Rutunda, Crystal Rugege, Gwydion Williams, Mira Emmanuel-Fabula, Alastair K. Denniston, Xiaoxuan Liu, Melissa Miles","doi":"10.1038/s41591-025-03815-3","DOIUrl":"https://doi.org/10.1038/s41591-025-03815-3","url":null,"abstract":"<p>Primary healthcare systems in sub-Saharan Africa face considerable challenges, particularly constrained clinical capacity¹ and variable quality-of-care delivery². Large language models (LLMs) represent a potentially transformative solution for clinical decision-making. Some LLMs have shown that they can recall clinical knowledge at expert level³ and can diagnose clinical vignettes (in simulated settings) more accurately than clinicians⁴. However, despite the fact that LLMs have shown promise, there is a striking lack of real-world evidence for their safe and effective use in clinical settings in Africa. Thus far, only handful of randomized trials of artificial intelligence (AI) for health (and none for generative AI tools) have been conducted on the continent⁵.</p><p>In Kenya, a pragmatic randomized controlled trial (RCT) has been embedded into routine care delivery undertaken at 16 of Penda Health’s clinics⁶, a Nairobi-based social enterprise that provides primary care services across all socioeconomic groups, from state social insurance recipients to privately insured and out-of-pocket payers. The trial aims to enroll 9,000 patients. The solution being trialed is an LLM-based ‘co-pilot’ feature that has been integrated directly into Penda’s electronic medical record system. Clinicians randomly assigned to the intervention arm receive automated, real-time suggestions for diagnosis, treatment planning and lab interpretation during patient consultations.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"76 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144546978","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Co-designing a public health data analytics platform","authors":"L. Otieno, P. Henderson, S. Khanna, N. Spurrier, L. J. Palmer, G. Hendrie, R. Mahoney, P. Arbon, J. Ferguson, P. Sharpe, M. Mittinty, H. Haji Ali Afzali, J. Rathjen, L. Bierbaum, G. Sallows, S. Omodei-James, S. Dahia, C. Miller, B. Bonevski, C. Ryder","doi":"10.1038/s41591-025-03806-4","DOIUrl":"https://doi.org/10.1038/s41591-025-03806-4","url":null,"abstract":"<p>Pandemics such as COVID-19 reveal unique challenges to public health systems worldwide, particularly the need for real-time insights into the risk of infectious disease spread to guide public health prevention and containment efforts¹. During the COVID-19 pandemic, Australia reported the second lowest prevalence of SARS-CoV-2 infections per 100,000 people and the third lowest number of confirmed or suspected COVID-19 deaths per million population, relative to the 37 other countries of the Organisation for Economic Co-operation and Development². Although many factors contributed to this success, evidence-based decision-making and nationwide collaboration through the use of predictive modeling were key. This allowed policymakers to evaluate population risk and compare potential public health outcomes associated with various disease-control strategies³.</p><p>Other specific factors also had a part in ensuring Australia’s overall success. For example, the ability to maintain public confidence through timely sharing of information with the public was crucial in South Australia. To maintain this level of trust and degree of social license, relationships with local media were strengthened, and a substantial effort was made to ensure clarity of messaging. In addition, health officials built strong relationships with community groups, particularly with priority populations (such as culturally and linguistically diverse populations) through face-to-face meetings, where appropriate, and regular webinars.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"148 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144546976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Too easy to swallow","authors":"","doi":"10.1038/s41591-025-03796-3","DOIUrl":"https://doi.org/10.1038/s41591-025-03796-3","url":null,"abstract":"Campaigners want ultra-processed foods to be better defined and regulated. But does the evidence support strong policy changes?","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"648 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144533157","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AAV gene therapy for autosomal recessive deafness 9: a single-arm trial","authors":"Jieyu Qi, Liyan Zhang, Ling Lu, Fangzhi Tan, Cheng Cheng, Yicheng Lu, Wenxiu Dong, Yinyi Zhou, Xiaolong Fu, Lulu Jiang, Chang Tan, Shanzhong Zhang, Sijie Sun, Huaien Song, Maoli Duan, Dingjun Zha, Yu Sun, Xia Gao, Lei Xu, Fan-Gang Zeng, Renjie Chai","doi":"10.1038/s41591-025-03773-w","DOIUrl":"https://doi.org/10.1038/s41591-025-03773-w","url":null,"abstract":"<p>Gene therapy for congenital deafness has shown promising results in children but lacks data in older populations. We conducted a single-arm trial of adeno-associated virus (AAV)-<i>O</i><i>TOF</i> gene therapy using the Anc80L65 capsid in ten participants with autosomal recessive deafness 9 aged 1.5 to 23.9 years at five sites in China. The primary endpoints were safety and tolerability within 5 years, and secondary endpoints assessed auditory function. Initial findings from the ten patients with 6–12 months of follow-up, including one patient who received two injections, revealed that the therapy was well tolerated, with 162 grade I/II adverse events. Decreased neutrophil percentage was the most common event (16 of 162). All ten participants had at least 6 months of follow-up and improved their pure-tone-average hearing level from baseline 106 ± 9 (mean ± s.d.) to 52 ± 30 decibels (dB). Other secondary endpoints showed similar improvements, including the average click auditory brainstem response (ABR) threshold, the tone-burst ABR threshold and the auditory steady-state response (101 ± 1 to 48 ± 26 dB, 91 ± 4 to 57 ± 19 dB and 80 ± 14 to 64 ± 21 dB, respectively). Post hoc analyses were conducted to evaluate the timecourse and factors contributing to the hearing improvement. Therapeutic effect was rapid, taking 1 month to achieve most of the overall hearing improvement. On an individual level, click and tone-burst ABR thresholds, but not the auditory steady-state response, reliably predicted the behavioral pure-tone-average thresholds after 4 months (<i>R</i>² = 0.68, 0.73 and 0.17, respectively). An age-dependent therapeutic effect was observed, with optimal outcomes in 5- to 8-year-olds. These preliminary results show that AAV-<i>OTOF</i> was safe and well tolerated in patients ranging from toddlerhood to adulthood. The trial remains ongoing and requires extended follow-up to confirm the long-term safety and efficacy. ClinicalTrials.gov registration: NCT05901480.</p>","PeriodicalId":19037,"journal":{"name":"Nature Medicine","volume":"47 1","pages":""},"PeriodicalIF":82.9,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144533154","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}