Texture-enhanced thermography for joint inflammation detection using consumer-grade thermal cameras

Area of Science:

• Rheumatology
• Medical Imaging
• Biophysics

Background:

• Early synovitis detection is crucial for treat-to-target management in inflammatory arthritis.
• Current sensitive methods like MRI and ultrasound have limitations (cost, portability, operator dependence).
• Conventional infrared thermography (IRT) struggles with ambient variability and bilateral disease.

Purpose of the Study:

• To evaluate consumer-grade thermal cameras with texture analysis and differencing for detecting joint inflammation.
• To assess if this augmented IRT can differentiate inflammatory arthritis from other conditions.
• To overcome limitations of traditional IRT in bilateral disease.

Main Methods:

• Acquired thermal images of multiple joints from 239 participants (arthritis, OA, fibromyalgia, controls).
• Extracted texture features (entropy, skewness) alongside absolute temperature.
• Applied contralateral (L-R) and anterior-posterior (A-P) differencing for normalization.

Main Results:

• Contralateral differencing significantly improved detection of unilateral inflammation (e.g., knee d=2.15).
• Texture features with A-P differencing detected subclinical wrist inflammation better than absolute temperature.
• Thermographic signatures differentiated inflammatory arthritis from OA and fibromyalgia.

Conclusions:

• Texture-enhanced IRT with L-R and A-P differencing shows potential for identifying active and subclinical joint inflammation.
• This method addresses limitations of traditional IRT, especially in bilateral disease.
• Augmented IRT offers a promising, accessible screening tool for joint inflammation.