Affiliations 

  • 1 Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA
  • 2 Department of Computer Science, Stanford University, Stanford, CA, USA
  • 3 Department of Radiology/Pediatric Radiology, Stanford University, Palo Alto, CA, USA
  • 4 Department of Clinical Radiology, Great Ormond Street Hospital for Children, London, UK
  • 5 Saint Louis University School of Medicine, Saint Louis University, Saint Louis, MO, USA
  • 6 Central Radiology Institute, Johannes Kepler University Linz, Kepler University Hospital GmbH, Linz, Austria
  • 7 Division of Pediatric Radiology, Department of Radiology, Medical University of Graz, Graz, Austria
  • 8 Department of Pediatrics, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada
  • 9 Artificial Intelligence Committee, Colombian Association of Radiology, Medellín, Colombia
  • 10 Education and Research Department, Servicios Especiales de Salud Hospital Universitario de Caldas, Manizales, Colombia
  • 11 Department of Pediatrics and Neonatology, Servicios Especiales de Salud Hospital Universitario de Caldas, Manizales, Colombia
  • 12 Department of Diagnostic and Interventional Radiology, Eberhard Karls University, Tuebingen University Hospital, Tuebingen, Germany
  • 13 Medical Image and Data Analysis (MIDAS) Laboratory, University Hospital of Tübingen, Tübingen, Germany
  • 14 Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany
  • 15 Lab for AI in Medicine, Department of Diagnostic and Interventional Radiology, University Hospital RWTH Aachen, Aachen, Germany
  • 16 Institute for AI in Medicine (IKIM), University Medicine Essen, Essen, Germany
  • 17 Department of Radiology, RSUP Dr Sardjito, Faculty of Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia
  • 18 Thoracic and Vascular Surgery Research Center, Namazi Hospital, Shiraz University of Medical Sciences, Shiraz, Iran
  • 19 Pediatric Pulmonology Institute, Dana Children's Hospital, Tel Aviv Sourasky University Medical Center (Ichilov), Tel Aviv, Israel
  • 20 Department of Diagnostic and Interventional Radiology, Graduate School of Medicine, Osaka Metropolitan University, Osaka, Japan
  • 21 Department of Artificial Intelligence, Graduate School of Medicine, Osaka Metropolitan University, Osaka, Japan
  • 22 Department of Diagnostic Radiology, American University of Beirut, Beirut, Lebanon
  • 23 Department of Paediatrics, Sultan Ahmad Shah Medical Centre (SASMEC), International Islamic University Malaysia, Kuantan, Malaysia
  • 24 Department of Radiology, Sultan Ahmad Shah Medical Centre @IIUM, International Islamic University Malaysia, Kuantan, Malaysia
  • 25 Department of Neonatal Pediatrics, The Indus Hospital and Health Network, Karachi, Pakistan
  • 26 Department of Diagnostic Radiology, The Indus Hospital and Health Network, Karachi, Pakistan
  • 27 Department of Neurosurgery, Sidra Medicine, Doha, Qatar
  • 28 Department of Radiology, Sidra Medicine, Doha, Qatar
  • 29 Manager of Diagnostic Imaging Section, King Abdul-Aziz Medical City, Ministry of National Guard Health Affairs, Riyadh, Saudi Arabia
  • 30 Pediatric Radiology Consultant, King Abdullah Specialized Children Hospital, Riyadh, Saudi Arabia
  • 31 Department of Pediatrics, Faculty of Medical Sciences, University of Kragujevac, Kragujevac, Serbia
  • 32 Department of Surgery, Faculty of Medical Sciences, University of Kragujevac, Kragujevac, Serbia
  • 33 Department of Diagnostic Imaging, National University Hospital Singapore, Singapore, Singapore
  • 34 Department of Paediatrics, New Somerset Hospital, University of Cape Town, Cape Town, South Africa
  • 35 Department of Radiology, La Paz University Hospital, Madrid, Spain
  • 36 Neonatal Intensive Care Unit, Biomedical Research and Innovation Institute of Cádiz (INiBICA), Research Unit, Puerta del Mar University Hospital, Cádiz, Spain
  • 37 Department of Neonatology, Life Memorial Hospital, Titu Maiorescu Faculty of Medicine, Bucharest, Romania
  • 38 Department of Pediatrics, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
  • 39 Department of Family and Community Medicine, Tri-Service General Hospital, National Defense Medical University, Taipei, Taiwan
  • 40 Medical Technology Education Center, School of Medicine, College of Medicine, National Defense Medical University, Taipei, Taiwan
  • 41 Department of Radiology, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  • 42 Center for AI in Medicine, Research Affairs, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
  • 43 Department of Family Medicine, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand
  • 44 Ministry of Public Health, Maharaj Nakorn Chiang Mai Hospital, Chiang Mai, Thailand
  • 45 Department of Pediatrics, Division of Neonatology, Istanbul Training and Research Hospital, Istanbul, Turkey
  • 46 Department of Pediatrics, Division of Neonatology, Kirikkale University School of Medicine, Kirikkale, Turkey
  • 47 University of Colorado School of Medicine, Children's Hospital Colorado, Aurora, CO, USA
  • 48 Department of Radiology, Newark Beth Israel Medical Center, RWJ Barnabas Health, Newark, NJ, USA
  • 49 Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, USA
  • 50 Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA. pranav_rajpurkar@hms.harvard.edu
Nat Biomed Eng, 2026 Oct 13.
PMID: 42869977 DOI: 10.1038/s41551-026-01812-6

Abstract

Despite growing clinical adoption of artificial intelligence in radiology, fewer than one in five commercially available models are cleared for paediatric use. Existing systems are trained predominantly on adult data and generalize poorly to neonates, yet efficient strategies for repurposing them without data sharing or large-scale local data curation are lacking. Here we present an international retrospective study of neonatal chest radiographs from 30 neonatal intensive care units across 20 countries and territories, evaluating continual learning as a privacy-preserving approach to adapt a deep learning model for adult endotracheal tube assessment to neonatal care. Each hospital contributed 50 radiographs for sequential model retraining without centralized data pooling. Continual learning achieved lower anatomical landmark localization error (0.67 cm versus 1.29 cm for hospital-specific fine-tuning and 4.30 cm for adult-only models) at 29 of 30 sites, with improved detection of deep intubations at 27 of 30 sites. These findings demonstrate that privacy-preserving collaborative learning can address critical gaps in paediatric AI deployment.

* Title and MeSH Headings from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.