From:  Thermography and technology: transforming health insurance with smart diagnostics and fraud prevention

 Summary of thermography diagnostic sensitivities from key studies.

ConditionSensitivity range (%)Key metrics/ExamplesReferences
Breast cancer83–98Thermal asymmetry > 3.0°C[6, 11]
Diabetic foot ulcers (DFUs)89–9896.71% with AI; temperatures 29.7 ± 2.9°C diabetic vs. 26.7 ± 1.6°C healthy[7, 12]
Rheumatoid arthritis (RA)N/A (no direct sensitivity reported)Combined with ultrasound shows superior correlation with DAS28 [e.g., r = 0.393 for MAX (PD), p-value = 0.016][24]
Orthopedic injuries/WoundsUp to 98Fracture tracking and healing monitoring[9, 15, 18, 22]
Other (e.g., stroke, skin cancer)N/A (no direct sensitivity reported)Facial asymmetry for stroke (e.g., > 0.5°C abnormal; specific sensitivities 33–56% for Wallenberg syndrome); subtle temperature differences for skin cancer (e.g., Δ 2–4 K for melanoma)[14, 25]
Return to article view

Ranges reflect variations due to protocols and equipment; all p-values < 0.05 were reported [6, 7, 12]. For RA and other conditions, focus is on correlation and qualitative metrics rather than direct sensitivity, with significant associations (e.g., p < 0.05) as per [14, 24, 25]. AI: artificial intelligence; DAS28: Disease Activity Score 28, a validated index for RA disease activity discussed in the main text.