ArticleBMJ digital health & AI2025
Evaluating the environmental sustainability of AI in radiology: a systematic review of current practice.
Article in BMJ digital health & AI, 2025. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Not yet cited in PubMed.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
0 citing papers in PubMed.
No citing paper in PubMed yet.
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
7 authors.
Funding
No grant is acknowledged in the PubMed record.
Abstract
Objective: Data-heavy and energy-heavy artificial intelligence (AI) technologies are increasingly being applied in healthcare, particularly for clinical imaging, often without consideration of their environmental impacts. We aimed to assess current practice in considering and evaluating environmental sustainability (ES) impacts of AI-enabled clinical pathways in radiology. Methods and analysis: We searched MEDLINE and Embase on 5 November 2024 for quantitative clinical radiology studies which used a machine learning approach to aid in radiological diagnosis or intervention and discussed its ES impacts. We included peer-reviewed, English language studies published from 2015 onwards. Our primary outcome was any quantitative reporting of ES impacts, and our secondary outcome was any within-text qualitative discussion of ES impacts. For quantitative outcomes, we conducted synthesis without meta-analysis based on effect direction and size, with our secondary outcome synthesised narratively. Results: Of 4449 records screened, 18 met our inclusion criteria. Six (33.33%) reported quantitative ES outcomes and 15 (83.33%) included qualitative discussion of ES. When applied to the same tasks, algorithms designed to be 'lightweight' demonstrated from 2.19 to 17.15 times reduction in carbon emissions (median 7.81, 16 datapoints) and from 1.60 to 751.62 times reduction in energy consumption (median 3.22, 16 datapoints) compared with state-of-the-art alternatives, while maintaining similar or improved clinical performance. No quantitative studies compared ES outcomes for an AI-enabled pathway versus standard-of-care, and 75.00% of studies reporting only on our secondary outcome included just a single sentence on sustainability. Conclusion: Despite increasing concern about the climate impacts of AI, environmental outcomes are rarely considered within evaluations of AI technologies in clinical radiology. However, there are approaches available with smaller carbon footprints. To meet their stated aims on sustainability, funders and governance bodies should consider how to promote integration of environmental impact assessment into AI health research and evaluation. PROSPERO registration number: CRD42024601818.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.