A clinically aligned multimodal workflow framework for chronic wound assessment: An evidence-informed conceptual...

Area of Science:

• Medical Artificial Intelligence
• Digital Health
• Wound Care Technology

Background:

• Chronic wound assessment relies on accurate and timely data.
• Image-based artificial intelligence (AI) shows promise for improving wound evaluation.
• Existing AI applications lack a standardized, clinically integrated workflow.

Purpose of the Study:

• To create an evidence-based, clinically aligned multimodal workflow framework for chronic wound assessment.
• To identify patterns in image-based AI studies to inform the framework's design.

Main Methods:

• Conducted a structured evidence mapping and synthesis of 44 published studies on image-based AI for chronic wound assessment.
• Extracted data on wound types, imaging methods, AI task domains, model architectures, performance, and deployment.
• Identified cross-study patterns to construct a conceptual workflow framework.

Main Results:

• The evidence base primarily focused on diabetic foot ulcers and general chronic wounds, with less on other wound types.
• Camera-based imaging was common, while device-based and mobile acquisition were less represented.
• AI tasks focused on classification and segmentation, with emerging use of explainability methods and clinically meaningful interpretation tasks.

Conclusions:

• A four-stage workflow framework (localization, segmentation, interpretation, monitoring) was developed, supported by current wound-AI evidence.
• This framework provides a conceptual structure for future multimodal system development and real-world validation in wound care.