Summary
This scoping review systematically mapped externally validated machine learning models used in cancer patient care from 2014–2022, synthesising evidence from 56 eligible studies. Convolutional neural networks predominated and demonstrated high performance; most studies were retrospective and multi-institutional, primarily using image-based data. Clinical utility assessments involving 499 clinicians indicated that AI assistance improved clinician performance relative to standard clinical systems, though the review identifies substantial gaps in external validation and clinical utility reporting across the cancer informatics literature.
Regional applicability
The review synthesised international evidence from high-quality journals but does not specify geographic origin of included studies. Findings on machine learning performance and clinical utility are methodologically transferable to United Kingdom oncology practice, though implementation would depend on local data governance, clinical infrastructure, and NHS adoption pathways.
Key measures
Machine learning model performance metrics; external validation status; clinical utility assessment tools; clinician performance improvement; comparison to standard clinical systems
Outcomes reported
The scoping review quantified the performance and clinical utility of externally validated machine learning models in cancer patient care, examining relationships between model types, cancer types, and specific clinical decisions. Clinical utility was assessed through measurement of clinician performance improvement and comparison to standard clinical systems.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.