Cutaneous Thermography in Arthropathies: Quantitative Imaging, Machine Learning, and Clinical Translation

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Arthropathies pose a significant global health burden, necessitating accurate and timely diagnosis.
• Conventional imaging techniques (ultrasound, MRI) have limitations like operator dependence and high costs.
• Infrared thermography offers a non-invasive alternative for detecting thermal changes associated with arthropathic conditions.

Purpose of the Study:

• To review recent advancements in thermography-based biomarkers for arthropathies.
• To explore the role of AI, machine learning, and deep learning in thermographic analysis.
• To address challenges and requirements for clinical implementation of thermography in arthropathies.

Main Methods:

• Review of recent literature on thermography and AI in arthropathy diagnosis.
• Summary of thermography-based biomarkers (e.g., asymmetry indices, hotspot burden).
• Discussion of machine learning and deep learning applications for prediction and phenotyping.

Main Results:

• Thermography, particularly quantitative and AI-enhanced approaches, shows potential for early diagnosis and disease activity assessment.
• AI integration improves prediction, phenotyping, and multimodal data analysis.
• Identified key thermographic biomarkers and AI methodologies.

Conclusions:

• Thermography combined with AI is a promising adjunct for arthropathy diagnosis and management.
• Standardization of protocols, external validation, and transparent reporting are crucial for large-scale clinical adoption.
• Further research is needed to overcome challenges like measurement variability and domain shift.