Dyspnoea is described as one of the most distressing symptoms experienced by patients with advanced lung cancer, substantially reducing quality of life and complicating palliative care. Conventional approaches to assessment and management commonly depend on episodic patient self-report, which may fail to capture the fluctuating and multidimensional character of breathlessness. Emerging artificial intelligence (AI) technologies could support more continuous symptom assessment, enable remote monitoring, improve risk prediction and provide decision support to clinicians and caregivers. However, the types of AI technologies applied to dyspnoea in advanced lung cancer, the data inputs these tools use, the outcomes reported and the clinical or home-care contexts in which they are deployed are not yet well described. This scoping review aims to map the available evidence and characterise how AI has been applied to dyspnoea management in this population.
The review will identify and map empirical evidence on AI applications for dyspnoea management among patients with advanced lung cancer. It will extract and summarise information on study characteristics, the types of AI technologies used, data sources feeding the algorithms, dyspnoea-related outcomes measured, implementation contexts (for example inpatient, outpatient or home care) and any reported feasibility or performance metrics. The work is designed to clarify the current landscape, identify gaps in evidence and inform subsequent research or implementation efforts.
This scoping review will follow the Joanna Briggs Institute (JBI) methodology for scoping reviews. The protocol is prepared in line with Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P) where applicable. The completed review will be reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR).
A comprehensive search will be conducted across multiple bibliographic and domain-specific databases from database inception with no date restrictions. Databases to be searched include PubMed/MEDLINE, Embase (Ovid), CINAHL (EBSCO), Web of Science Core Collection, Scopus, Cochrane Library, IEEE Xplore, ACM Digital Library, PsycINFO, CNKI, WanFang Data, VIP Database and SinoMed. Supplementary search sources comprise Google Scholar, ClinicalTrials.gov, the WHO International Clinical Trials Registry Platform (ICTRP), medRxiv, arXiv, ProQuest Dissertations & Theses and reference lists of included studies and relevant reviews. The search strategy will be designed to capture terms related to dyspnoea, advanced lung cancer and AI technologies, and will be adapted to each platform.
Eligible evidence will include original quantitative, qualitative and mixed-methods studies that report empirical data related to AI applications for dyspnoea in advanced lung cancer. Full grey-literature reports with sufficient empirical data will also be considered. Systematic, scoping and narrative reviews will not be included as evidence units but will be used to identify additional primary studies. The protocol does not impose date restrictions. Any study or report that lacks empirical data sufficient for charting will be excluded.
Two reviewers will independently screen titles and abstracts, followed by full-text screening of potentially relevant records. Disagreements at any stage will be resolved through discussion and, if necessary, adjudication by a third reviewer. The process and reasons for exclusion at full-text review will be documented and reported in accordance with PRISMA-ScR guidance.
Data will be charted using a piloted extraction form. Extracted items will include study characteristics (for example authorship, setting and study design), AI technology type or algorithm class, data sources used to develop or feed the AI (for example physiological signals, patient-reported outcomes or imaging), dyspnoea-related outcomes reported, implementation context (clinical or home-care), and any reported feasibility, usability or performance metrics. If available, information on population characteristics and care pathways relevant to implementation will also be captured.
Extracted data will be synthesised descriptively and thematically to map the landscape of AI applications for dyspnoea management in advanced lung cancer. Descriptive synthesis will summarise the distribution of study designs, AI technologies, data inputs and outcome types. Thematic synthesis will identify common implementation contexts, reported benefits or limitations, and gaps in evidence. Where studies report feasibility or performance metrics, these will be summarised narratively; the protocol does not plan meta-analysis.
Because the review uses published and publicly available literature and does not involve primary patient data, ethical approval is not required. Findings will be disseminated through a peer-reviewed results paper, conference presentations and summary briefs targeted to clinicians, researchers and policymakers. The intention is to inform future research priorities and potential clinical or home-care implementation of AI tools for dyspnoea management.
This scoping review protocol has been registered in the International Prospective Register of Systematic Reviews (PROSPERO) with registration number CRD420251112279.