Feasibility and utility of infrared thermography to distinguish between acute presentations of cellulitis and deep venous thrombosis in the emergency department

Abstract

Study objectives: Specific illnesses may manifest with unique regional heat variations and patterns in body surface temperature that can be detected by infrared thermography. Although there is some experience described in the literature with infrared thermography in burns and cancer, there is little experience with infrared thermography in other disease states. This study attempted to determine the feasibility and utility of infrared thermography to distinguish between lower-extremity cellulitis and deep venous thrombosis. Methods: This is a case-control study of infrared thermography using a ThermaCAM P60 device (FLIR Systems, North Billerica, MA). Infrared thermography images were obtained before diagnostic duplex imaging in 2 patients presenting to the emergency department of a tertiary hospital with acute pain, swelling, and erythema of a leg. The camera has a spatial resolution of 320×240 pixels in focal plane array, scan integration rate of 60 Hz, and thermal sensitivity of 0.08^oC per pixel. Focal distance was 1 meter, with a set emissivity of 0.98. Images are acquired in JPEG format. Body surface temperature in each extremity was analyzed using specialized software across an area 4 cm², placed over the area of interest. The mean temperature change in the combined left-right (dT/dx), and up-down (dT/dy) directions in the affected versus normal extremities were calculated using the formula:<mml:math><mml:mrow><mml:mrow><mml:mo>∑</mml:mo><mml:mo>(</mml:mo><mml:mo>|</mml:mo><mml:mi>dT</mml:mi><mml:mi>dx</mml:mi><mml:mo>|</mml:mo><mml:mo>+</mml:mo><mml:mo>|</mml:mo><mml:mi>dT</mml:mi><mml:mi>dy</mml:mi><mml:mo>|</mml:mo><mml:mo>)</mml:mo></mml:mrow><mml:mi>N</mml:mi></mml:mrow></mml:math> where N equals the number of pixels in the analyzed area. Results: For the patient with cellulitis of the leg, the maximum dT/dx and dT/dy values for the affected leg were 0.68 and 0.56, respectively, versus 0.19 and 0.22 in the normal extremity. The mean temperature change in the combined directions was 0.35 in the affected extremity versus 0.09 in the normal extremity. For the patient with deep venous thrombosis of the leg, the maximum dT/dx and dT/dy values for the affected leg were 0.55 and 0.34, respectively, versus 0.26 and 0.17 in the normal extremity. The mean temperature change in the combined directions was 0.14 in the affected extremity versus 0.09 in the normal extremity. Conclusion: The infrared thermography images of an extremity affected by cellulitis versus deep venous thrombosis are visually quite different. Mathematical models of these 2 clinical entities using infrared thermography also appear to present distinct results. Because infrared thermography is easily and rapidly obtained and poses no risk or radiation to the patient, this technology may prove useful in distinguishing between cellulitis and deep venous thrombosis in the clinical setting. Further study is required to determine the operating characteristics of the infrared thermography test.