Pub Date : 2026-01-26DOI: 10.1038/s41551-025-01586-3
{"title":"AI learns across species to address human clinical imaging data sparsity","authors":"","doi":"10.1038/s41551-025-01586-3","DOIUrl":"https://doi.org/10.1038/s41551-025-01586-3","url":null,"abstract":"","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"1 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048410","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 : 2026-01-26DOI: 10.1038/s41551-025-01605-3
Hedan Bai, Jianlin Zhou, Mingzheng Wu, Steven Papastefan, Xiuyuan Li, Haohui Zhang, Kaiyu Zhao, Zhuoran Zhang, Wei Ouyang, Catherine R. Redden, Amir M. Alhajjat, Heyang Wang, Yibo Zhou, Kenneth Madsen, Shuo Li, Andrew I. Efimov, Katelyn Ma, Lisa Kovacs, Sahdev Patel, Daniel R. Liesman, Katherine C. Ott, Rinaldo Garziera, Steffen Sammet, Wenming Zhang, Yonggang Huang, Aimen F. Shaaban, John A. Rogers
{"title":"A filamentary soft robotic probe for multimodal in utero monitoring of fetal health","authors":"Hedan Bai, Jianlin Zhou, Mingzheng Wu, Steven Papastefan, Xiuyuan Li, Haohui Zhang, Kaiyu Zhao, Zhuoran Zhang, Wei Ouyang, Catherine R. Redden, Amir M. Alhajjat, Heyang Wang, Yibo Zhou, Kenneth Madsen, Shuo Li, Andrew I. Efimov, Katelyn Ma, Lisa Kovacs, Sahdev Patel, Daniel R. Liesman, Katherine C. Ott, Rinaldo Garziera, Steffen Sammet, Wenming Zhang, Yonggang Huang, Aimen F. Shaaban, John A. Rogers","doi":"10.1038/s41551-025-01605-3","DOIUrl":"https://doi.org/10.1038/s41551-025-01605-3","url":null,"abstract":"","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"12 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048411","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 : 2026-01-23DOI: 10.1038/s41551-025-01553-y
Brendan C Jones,Giada Benedetti,Giuseppe Calà,Ramin Amiri,Lucinda Tullie,Roberto Lutman,Jahangir Sufi,Lucy Holland,Daniyal J Jafree,Monika Balys,Glenn Anderson,Ian C Simcock,Owen J Arthurs,Simon Eaton,Nicola Elvassore,Vivian Sw Li,Christopher J Tape,Kelsey Dj Jones,Camilla Luni,Giovanni Giuseppe Giobbe,Paolo De Coppi
Patient-derived human organoids have the capacity to self-organize into more complex structures. However, to what extent gastric organoids can recapitulate differentiated cell types and mucosal functions remains unexplored. Here we report on how region-specific gastric organoids can self-assemble into complex multi-regional assembloids. These assembloids show increased complexity and cross-communication between different gastric regions, allowing for the emergence of the elusive parietal cell type that is responsible for the production of gastric acid and shows a functional response to drugs targeting the H+/K+ ATPase pump. We generate assembloids from paediatric patients with a genetic condition found to be associated with unusual antral foveolar hyperplasia and hyperplastic polyposis. Our multi-regional assembloid efficiently recapitulates hyperplastic-like antral regions, with decreased mucin secretion and glycosylated H+/K+ ATPase subunit beta, which results in impaired gastric acid secretion. Multi-regional gastric assembloids, generated using paediatric-stem-cell-derived organoids, successfully recapitulate the structural and functional characteristics of the human stomach, offering a promising tool for studying gastric epithelial interactions and disease mechanisms that were previously challenging to investigate in primary models.
{"title":"Human gastric multi-regional assembloids for functional parietal maturation and patient-specific modelling of antral foveolar hyperplasia.","authors":"Brendan C Jones,Giada Benedetti,Giuseppe Calà,Ramin Amiri,Lucinda Tullie,Roberto Lutman,Jahangir Sufi,Lucy Holland,Daniyal J Jafree,Monika Balys,Glenn Anderson,Ian C Simcock,Owen J Arthurs,Simon Eaton,Nicola Elvassore,Vivian Sw Li,Christopher J Tape,Kelsey Dj Jones,Camilla Luni,Giovanni Giuseppe Giobbe,Paolo De Coppi","doi":"10.1038/s41551-025-01553-y","DOIUrl":"https://doi.org/10.1038/s41551-025-01553-y","url":null,"abstract":"Patient-derived human organoids have the capacity to self-organize into more complex structures. However, to what extent gastric organoids can recapitulate differentiated cell types and mucosal functions remains unexplored. Here we report on how region-specific gastric organoids can self-assemble into complex multi-regional assembloids. These assembloids show increased complexity and cross-communication between different gastric regions, allowing for the emergence of the elusive parietal cell type that is responsible for the production of gastric acid and shows a functional response to drugs targeting the H+/K+ ATPase pump. We generate assembloids from paediatric patients with a genetic condition found to be associated with unusual antral foveolar hyperplasia and hyperplastic polyposis. Our multi-regional assembloid efficiently recapitulates hyperplastic-like antral regions, with decreased mucin secretion and glycosylated H+/K+ ATPase subunit beta, which results in impaired gastric acid secretion. Multi-regional gastric assembloids, generated using paediatric-stem-cell-derived organoids, successfully recapitulate the structural and functional characteristics of the human stomach, offering a promising tool for studying gastric epithelial interactions and disease mechanisms that were previously challenging to investigate in primary models.","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"29 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033854","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 : 2026-01-22DOI: 10.1038/s41551-026-01611-z
It is an exciting time for biomedical engineering, with advances rapidly reshaping the forefront of translational research and medicine. Here we highlight some areas and technologies that we are particularly excited about for the coming years.
{"title":"On the horizon in biomedical engineering","authors":"","doi":"10.1038/s41551-026-01611-z","DOIUrl":"10.1038/s41551-026-01611-z","url":null,"abstract":"It is an exciting time for biomedical engineering, with advances rapidly reshaping the forefront of translational research and medicine. Here we highlight some areas and technologies that we are particularly excited about for the coming years.","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"10 1","pages":"1-2"},"PeriodicalIF":26.8,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s41551-026-01611-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146016503","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}
Pub Date : 2026-01-22DOI: 10.1038/s41551-025-01587-2
Zifeng Wang,Benjamin Danek,Ziwei Yang,Zheng Chen,Jimeng Sun
Large language models (LLMs) can generate impressive data visualizations from simple requests, yet their accuracy remains underexplored. Here we present a benchmark of 293 coding tasks derived from 39 studies across 7 biomedical research areas, including biomarkers, integrative analysis, genomic profiling, molecular characterization, therapeutic response, translational research and pan-cancer analysis. Benchmarking eight proprietary and eight open-source LLMs under various prompting strategies reveals an overall accuracy below 40%. This low accuracy raises serious concerns about the risk of propagating incorrect scientific findings when blindly relying on AI-generated analyses. Therefore, we develop an AI agent that begins with and iteratively refines an analysis plan before generating code, achieving 74% accuracy. We embody this insight in a platform that enables users to codevelop analysis plans with LLMs and execute them within an integrated environment. In a user study with five medical researchers, the platform enabled users to complete over 80% of the analysis code for three studies.
{"title":"Making large language models reliable data science programming copilots for biomedical research.","authors":"Zifeng Wang,Benjamin Danek,Ziwei Yang,Zheng Chen,Jimeng Sun","doi":"10.1038/s41551-025-01587-2","DOIUrl":"https://doi.org/10.1038/s41551-025-01587-2","url":null,"abstract":"Large language models (LLMs) can generate impressive data visualizations from simple requests, yet their accuracy remains underexplored. Here we present a benchmark of 293 coding tasks derived from 39 studies across 7 biomedical research areas, including biomarkers, integrative analysis, genomic profiling, molecular characterization, therapeutic response, translational research and pan-cancer analysis. Benchmarking eight proprietary and eight open-source LLMs under various prompting strategies reveals an overall accuracy below 40%. This low accuracy raises serious concerns about the risk of propagating incorrect scientific findings when blindly relying on AI-generated analyses. Therefore, we develop an AI agent that begins with and iteratively refines an analysis plan before generating code, achieving 74% accuracy. We embody this insight in a platform that enables users to codevelop analysis plans with LLMs and execute them within an integrated environment. In a user study with five medical researchers, the platform enabled users to complete over 80% of the analysis code for three studies.","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"58 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146021521","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 : 2026-01-21DOI: 10.1038/s41551-025-01594-3
Lukas Ehlen, Martí Farrera-Sal, Martin Szyska, Janine Arndt, Simon Schallenberg, Cedric Scholz, Mingxing Yang, Claudia Vollbrecht, Anna Löwa, Rebecca Friedrich, Marco Mai, Lena Peter, Samira Picht, Sarah Schulenberg, Daniel Geray, Gabriela Korus, Anke Sommerfeld, Denise Treue, Julia Strauchmann, Aron Elsner, Jonas Kath, Valeria Fernandez Vallone, Maria Joosten, Franka Klatte-Schulz, Ansgar Petersen, Harald Stachelscheid, Dimitrios L. Wagner, Claudia Spies, Jens-Carsten Rückert, Andreas C. Hocke, Julia K. Polansky, Regina Stark, Oliver Klein, Michael Schmueck-Henneresse
Lung cancer, the leading cause of cancer-related mortality, presents major challenges for both standard therapies and chimeric antigen receptor (CAR) T cell therapy due to tumour heterogeneity and resistance. Preclinical models that capture patient-specific factors are essential for personalizing treatment decisions. Here we show that matched lung tumouroids and healthy lung organoids derived from patients provide a robust platform for studying therapy responses. The tumouroids faithfully retained the molecular and histological identity of the original tumours, as confirmed by genomic, epigenomic and proteomic analyses, and accurately replicated individual patient responses to standard-of-care therapies. Importantly, the platform also revealed patient-specific CAR T cell responses, uncovering a complex interplay between target antigen density and broader, tumour-intrinsic resistance programmes. By capturing these individualized factors, our model supports rational patient selection for CAR T cell therapy in lung cancer and provides a framework for designing CAR T cells tailored to overcome resistance mechanisms in solid tumours.
{"title":"Lung tumouroids as a testing platform for precision CAR T cell therapy","authors":"Lukas Ehlen, Martí Farrera-Sal, Martin Szyska, Janine Arndt, Simon Schallenberg, Cedric Scholz, Mingxing Yang, Claudia Vollbrecht, Anna Löwa, Rebecca Friedrich, Marco Mai, Lena Peter, Samira Picht, Sarah Schulenberg, Daniel Geray, Gabriela Korus, Anke Sommerfeld, Denise Treue, Julia Strauchmann, Aron Elsner, Jonas Kath, Valeria Fernandez Vallone, Maria Joosten, Franka Klatte-Schulz, Ansgar Petersen, Harald Stachelscheid, Dimitrios L. Wagner, Claudia Spies, Jens-Carsten Rückert, Andreas C. Hocke, Julia K. Polansky, Regina Stark, Oliver Klein, Michael Schmueck-Henneresse","doi":"10.1038/s41551-025-01594-3","DOIUrl":"https://doi.org/10.1038/s41551-025-01594-3","url":null,"abstract":"Lung cancer, the leading cause of cancer-related mortality, presents major challenges for both standard therapies and chimeric antigen receptor (CAR) T cell therapy due to tumour heterogeneity and resistance. Preclinical models that capture patient-specific factors are essential for personalizing treatment decisions. Here we show that matched lung tumouroids and healthy lung organoids derived from patients provide a robust platform for studying therapy responses. The tumouroids faithfully retained the molecular and histological identity of the original tumours, as confirmed by genomic, epigenomic and proteomic analyses, and accurately replicated individual patient responses to standard-of-care therapies. Importantly, the platform also revealed patient-specific CAR T cell responses, uncovering a complex interplay between target antigen density and broader, tumour-intrinsic resistance programmes. By capturing these individualized factors, our model supports rational patient selection for CAR T cell therapy in lung cancer and provides a framework for designing CAR T cells tailored to overcome resistance mechanisms in solid tumours.","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"48 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146006079","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}
Successful surgical resection of solid tumours requires highly reliable real-time intraoperative tools to accurately delineate tumour boundaries, which remains challenging in routine clinical standards. Here, we identify endogenous substances with intense autofluorescence in the second near-infrared window (NIR-II, 1,000-1,700 nm) that are abundant in human liver tissues but negligible in cancerous tissues. Inspired by this discovery, we develop a label-free and wide-field imaging approach, named tissue autofluorescence NIR-II imaging (TANI) for visualizing human liver malignancies. TANI demonstrates exceptional contrast (7.69 ± 0.52), sensitivity (97.8%) and specificity (98.4%) in delineating various liver malignancies, including hepatocellular carcinoma, intrahepatic cholangiocarcinoma and liver metastasis from cirrhotic or non-cirrhotic livers, outperforming routine fluorescence-guided surgery and conventional autofluorescence imaging in the visible (400-650 nm) or first near-infrared (700-900 nm) window. The excellent performance of TANI remains unaffected by cancer grade/stage, benign lesions or blood/bile contamination. These findings represent a promising advance in intraoperative decision-making and suggest a strong correlation between near-infrared autofluorescence and diseases. We believe that clarifying the molecular insights underlying these autofluorescent substances may provide new diagnostic directions.
{"title":"Label-free tissue NIR-II autofluorescence imaging for visualization of human liver malignancy.","authors":"Haisheng He,Wenwei Zhu,Han Miao,Shangfeng Wang,Zunguo Du,Hongxin Zhang,Jiang Ming,Ben Shi,Hao Wang,Jianping Qi,Yong Fan,Wei Wu,Dongyuan Zhao,Lun-Xiu Qin,Fan Zhang","doi":"10.1038/s41551-025-01593-4","DOIUrl":"https://doi.org/10.1038/s41551-025-01593-4","url":null,"abstract":"Successful surgical resection of solid tumours requires highly reliable real-time intraoperative tools to accurately delineate tumour boundaries, which remains challenging in routine clinical standards. Here, we identify endogenous substances with intense autofluorescence in the second near-infrared window (NIR-II, 1,000-1,700 nm) that are abundant in human liver tissues but negligible in cancerous tissues. Inspired by this discovery, we develop a label-free and wide-field imaging approach, named tissue autofluorescence NIR-II imaging (TANI) for visualizing human liver malignancies. TANI demonstrates exceptional contrast (7.69 ± 0.52), sensitivity (97.8%) and specificity (98.4%) in delineating various liver malignancies, including hepatocellular carcinoma, intrahepatic cholangiocarcinoma and liver metastasis from cirrhotic or non-cirrhotic livers, outperforming routine fluorescence-guided surgery and conventional autofluorescence imaging in the visible (400-650 nm) or first near-infrared (700-900 nm) window. The excellent performance of TANI remains unaffected by cancer grade/stage, benign lesions or blood/bile contamination. These findings represent a promising advance in intraoperative decision-making and suggest a strong correlation between near-infrared autofluorescence and diseases. We believe that clarifying the molecular insights underlying these autofluorescent substances may provide new diagnostic directions.","PeriodicalId":19063,"journal":{"name":"Nature Biomedical Engineering","volume":"63 1","pages":""},"PeriodicalIF":28.1,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146005415","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}
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：