Nature biomedical engineering · 2026

Repurposing adult AI radiograph models for neonatal care through continual learning: an international multicentre study.

СтатьяНазвание и аннотация: Перевод пока не подготовлен

Обработка статьи · этапы, ошибки и управление

Загрузка состояния…

История · было → стало

Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.

Rao Vishwanatha M, Chen Emma, Saenz Agustina, Banerjee Oishi, Marklund Henrik, Bijnens Lucas, Kim Sung Eun, Zucker Evan, Shelmerdine Susan, Palacio Jeremy, Kleinschmidt Kent Ryan, Suryadevara Sathvik, Jaliparthi Sri Sai Dinesh, Lu Jiaxi, Akula Rithvik, Scala Mario, Niedermair Julian F, Stranger Nikolaus, Tschauner Sebastian, Schmölzer Georg M, Veloza Vega Maria Jose, Osorio Valencia Juan Sebastián, Robledo Oscar Jaramillo, López Uribe Oscar Julian, Afat Saif, Küstner Thomas, Tayebi Arasteh Soroosh, Truhn Daniel, Pomykala Kelsey L, Kleesiek Jens, Afifah Naela Himayati, Irfan Rozan M, Shahriarirad Reza, Erfani Amirhossein, Amirav Israel, Be'er Moria, Mitsuyama Yasuhito, Ueda Daiju, Abi-Ghanem Alain S, Al Zakleet Safaa, Mohamed Mossad Abdelhak Shaban, Bin W Mustapha Wan Irfan, Nizami Naila, Khan Rafeah, Al-Kharazi Khalid, Iqbal Zafar, Al Shahrani Abdullah, Kashgari Amna, Prodanovic Tijana, Prodanovic Nikola, Lim Desmond Shi Wei, Hallinan James Thomas Patrick Decourcy, Nongena Phumza, Cooke Melissa Louise, Bret-Zurita Montserrat, Alonso-Ojembarrena Almudena, Nemes Alexandra, Lin Yung-Chieh, Lin Wei-Ting, Fang Wen-Hui, Lin Chin, Mahayosnond Atchara, Sriswasdi Sira, Khumrin Piyapong, Tanmuk Suphakris, Arman Didem, Cömert Serdar, Alan Serdar, Tandircioglu Umit Ayse, Mong David Andrew, Chung Esther, Sadler Michael, McBee Morgan P, Jones Richard H, Rajpurkar Pranav

Аннотация

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.

Repurposing adult AI radiograph models for neonatal care through continual learning: an international multicentre study.

Ошибки и ограничения

Аннотация