Impact of skin tone, environmental, and technical factors on thermal imaging

Abstract

Background Erythema, an early visual indicator of tissue damage preceding pressure injuries (PrIs), presents as redness in light skin tones but is harder to detect in dark skin tones. While thermography shows promise for early PrI detection, validation across different skin tones remains limited. Furthermore, most protocols and models have been developed under highly controlled conditions. Objective To evaluate how environmental and technical factors (i.e., patient positioning, lighting, distance, camera type) and skin tone impact thermal imaging measurements and temperature change. Methods This pre-post experimental study enrolled 35 healthy adults (30 with Monk Skin Tone Scale ≥6). Melanin Index was measured on the volar forearm using the SkinColorCatch ® . After baseline imaging, a 15.5°C cooling stone was placed on one posterior superior iliac spine (PSIS) for 5 minutes. Thermal images were then collected with either the FLIR E8-XT or the FLIR ONE Pro camera under varied conditions: two lighting types (overhead room versus localized LED ring light), three postures (side-lying, side-lying with forward knee placement, and side-lying with rearward knee placement), and two camera-to-body distances (35 cm and 50 cm from the PSIS). The cooling/imaging procedure was repeated using the alternate camera, and data were analyzed using mixed-methods model. Results Temperature change was effectively detected across all skin tones, with cooling resulting in a −3.7 ± 1.2°C difference between the region of interest (ROI) and control region. Camera type significantly affected measurements, with the ONE Pro recording 1.04°C less cooling than the E8-XT. Distance had minimal impact (0.11°C cooler at 50 cm vs 35 cm at baseline), with no significant difference when comparing ROI to control regions, while lighting and posture had no impact on measurements. Skin tone varied by cooling measurement, with higher melanin levels showing greater temperature changes. A 0.98°C difference was observed between the lightest and darkest skin tone groups. Discussion Our findings confirm thermal imaging’s robustness across varied environments, with the minor distance effects mitigated through perpendicular measurements and relative temperature comparisons. Significant discrepancies between thermal cameras (>1°C) highlight that these technologies cannot be used interchangeably when establishing thresholds. While effective across all skin tones, the observed differences in cooling response suggest tailored thresholds may be necessary for darker skin tones. Future research should focus on clinical validation across diverse populations to enhance PrI detection accuracy.