Considering the missing science of retraining and maintenance in medical artificial intelligence, using ophthalmology as an exemplar

Considerations around model retraining are standard practice in industry and non-healthcare sectors; however, this is much less well explored in medical artificial intelligence (AI). The problem is not only that models often fail to generalise, but that academia in particular does not have a systematic science of retraining to address this gap. This matters for building trustworthy models capable of making a lasting impact, rather than compounding as research waste. In this Perspective, we highlight three common challenges that constrain model retraining in medicine, and argue that academia must evolve beyond a focus on developing proofs-of-concept and world-first innovations to also recognise model retraining as scholarship. Drawing from case examples in ophthalmology, we call on stakeholders to consider not just how we build AI models, but how we should retrain, maintain, and share them.

Authors

Institutions

Publication Details

Journal
npj Digital Medicine
Published
2026-02-25
DOI
https://doi.org/10.1038/s41746-026-02463-2
Citations
1
Primary Topic
Retinal Imaging and Analysis
Type
article
Field-Weighted Citation Impact
31.92

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Considering the missing science of retraining and maintenance in medical artificial intelligence, using ophthalmology as an exemplar

Inês Laíns, Justin Engelmann, David A. Merle, Ariel Yuhan Ong et al.
1 citations
npj Digital Medicine
Retinal Imaging and Analysis
31.92
article

Considering the missing science of retraining and maintenance in medical artificial intelligence, using ophthalmology as an exemplar

Inês Laíns, Justin Engelmann, David A. Merle, Ariel Yuhan Ong, Yukun Zhou, Hyunmin Kim, Pearse A. Keane, Robbert R. Struyven, Alastair K. Denniston
article en
1 citations

Abstract

Considerations around model retraining are standard practice in industry and non-healthcare sectors; however, this is much less well explored in medical artificial intelligence (AI). The problem is not only that models often fail to generalise, but that academia in particular does not have a systematic science of retraining to address this gap. This matters for building trustworthy models capable of making a lasting impact, rather than compounding as research waste. In this Perspective, we highlight three common challenges that constrain model retraining in medicine, and argue that academia must evolve beyond a focus on developing proofs-of-concept and world-first innovations to also recognise model retraining as scholarship. Drawing from case examples in ophthalmology, we call on stakeholders to consider not just how we build AI models, but how we should retrain, maintain, and share them.

npj Digital Medicine
Massachusetts Eye and Ear Infirmary (US), University College Hospital (GB), Moorfields Eye Hospital NHS Foundation Trust (GB), Harvard University (US), NIHR Moorfields Biomedical Research Centre (GB), Centre for Innovation in Regulatory Science (GB), Moorfields Eye Hospital (GB), NIHR Birmingham Biomedical Research Centre (GB), University College London (GB), University of Birmingham (GB)
Massachusetts Lions Eye Research Fund, Wellcome Trust, National Institute for Health and Care Research, Ministerio de Economía y Competitividad, Moorfields Eye Charity, National Institutes of Health
Industry, innovation and infrastructure
Openalex Percentile: Top 1%
Retinal Imaging and Analysis
31.92
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.