Arterial intelligence: laying the foundations for holistic, multi-modal, and ai-driven clinical decision support in peripheral arterial disease
File(s)
Author(s)
Bergman, Henry
Type
Thesis
Abstract
Peripheral Arterial Disease (PAD) is a major global health burden, causing substantial morbidity, limb loss, and mortality. Although vascular diagnostics and treatments have advanced, decision-making in PAD remains variable and highly dependent on clinician experience, subjective assessment, and fragmented data sources. As clinical data grow in volume and complexity, artificial intelligence (AI) offers a powerful opportunity to improve risk stratification, enhance decision-making, and support more consistent, precise care.
This thesis establishes the foundations for a holistic, multi-modal, AI-driven Clinical Decision Support System (CDSS) for PAD. The work is organised into three components:
System requirements and conceptual foundations: A systematic review examines how vascular clinical decisions are currently made, analysing assessment frameworks and identifying key gaps and fragmentation that limit holistic care. These insights define the requirements of an ideal CDSS.
AI-enhanced exploration of individual data modalities: A suite of computational models is developed to evaluate core concepts identified in the system specification. These include structured questionnaire data, audio recordings, interview transcripts, electronic health records, computational imaging, and clinical photographs. Each modality is examined for its unique contribution to a unified decision-support architecture.
Roadmap for future development: The thesis reflects on the strengths and limitations of the proposed framework and outlines steps for further development, validation, integration, and clinical translation. It also discusses the theoretical principles underlying the integration of high-dimensional, multimodal data for structured, clinically meaningful outputs.
Explainability, ethics, and clinical usability are emphasised throughout, addressing transparency, regulatory considerations, and the safe, equitable deployment of AI tools in healthcare.
Overall, this research lays the groundwork for a standardized, data-driven, and patient-centred approach to PAD management, with the potential to improve diagnostic consistency, guide optimal treatment strategies, and enhance patient outcomes.
This thesis establishes the foundations for a holistic, multi-modal, AI-driven Clinical Decision Support System (CDSS) for PAD. The work is organised into three components:
System requirements and conceptual foundations: A systematic review examines how vascular clinical decisions are currently made, analysing assessment frameworks and identifying key gaps and fragmentation that limit holistic care. These insights define the requirements of an ideal CDSS.
AI-enhanced exploration of individual data modalities: A suite of computational models is developed to evaluate core concepts identified in the system specification. These include structured questionnaire data, audio recordings, interview transcripts, electronic health records, computational imaging, and clinical photographs. Each modality is examined for its unique contribution to a unified decision-support architecture.
Roadmap for future development: The thesis reflects on the strengths and limitations of the proposed framework and outlines steps for further development, validation, integration, and clinical translation. It also discusses the theoretical principles underlying the integration of high-dimensional, multimodal data for structured, clinically meaningful outputs.
Explainability, ethics, and clinical usability are emphasised throughout, addressing transparency, regulatory considerations, and the safe, equitable deployment of AI tools in healthcare.
Overall, this research lays the groundwork for a standardized, data-driven, and patient-centred approach to PAD management, with the potential to improve diagnostic consistency, guide optimal treatment strategies, and enhance patient outcomes.
Version
Open Access
Date Issued
2025-07-23
Date Awarded
01/01/2026
License URL
Advisor
Davies, Alun
Publisher Department
Department of Surgery & Cancer
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)