Thermal imaging could aid diagnosis in cellulitis, pseudocellulitis

Key takeaways:

• Overdiagnosis of cellulitis can lead to public health and safety risks and avoidable health care costs.
• Thermal imaging found significant temperature differences in skin with cellulitis vs. pseudocellulitis.

The use of thermal imaging and the ALT-70 prediction model were helpful in differentiating cellulitis from pseudocellulitis, reducing misdiagnosis, according to a study.

Cellulitis, characterized by redness, warmth, tenderness and edema, is responsible for nearly 3% of all ED visits; however, these symptoms are similar to many other conditions, collectively classified as pseudocellulitis. These include venous stasis dermatitis, deep venous thrombosis, drug eruptions, lymphedema and gout.

“One report found that approximately 30% of patients admitted to the hospital from the ED with a diagnosis of cellulitis actually have pseudocellulitis,” Michael S. Pulia, MD, PhD, associate professor at University of Wisconsin-Madison School of Medicine and Public Health, and colleagues wrote. “These diagnostic errors represent a public health and patient safety risk associated with unnecessary antibiotic prescribing (eg, bacterial resistance and serious adverse drug events) and are estimated to generate $515 million in avoidable health care expenditures annually.”

A prospective diagnostic validation study included 204 patients (mean age, 56.6 years; 59.3% men; 88.2% white) who presented with lower extremity symptoms at one emergency center from Oct. 11, 2018, through March 11, 2020.

Photographs and skin surface temperature measurements were taken from affected and unaffected skin in each patient.

Using a thermal camera (FLIR One Pro, Teledyne FLIR LLC) attached to an iPad (Apple), surface skin temperatures were taken from affected skin and the corresponding unaffected skin on the contralateral leg. A maximum and mean temperature of both locations was then generated using the thermal images and the camera’s software.

Of the 204 subjects, an independent panel of six board-certified physicians found that 92 (45.1%) had cellulitis.

A mean maximum temperature of 33.2°C (95% CI, 32.8°C-33.6°C) was recorded in the patients with cellulitis, whereas those with pseudocellulitis had a mean maximum temperature of 31.2°C (95% CI, 30.7°C -31.7°C), which was a statistically significant difference of 2°C (95% CI, 1.3°C-2.7°C; P < .001).

A maximum temperature difference between affected and unaffected skin was found in both cellulitis and pseudocellulitis patients. In the cellulitis group a difference of 2.9°C (95% CI, 2.5°C-3.4°C; P < .001) was found, whereas a difference of 0.9°C (95% CI, 0.5°C-1.3°C; P < .001) was found in the pseudocellulitis group.

Previous studies have used both thermal imaging and the ALT-70 — or asymmetry, leukocytosis, tachycardia and age of 70 years or older — model to predict cellulitis.

In this study, ALT-70 scores were determined between 0 and 7, with a score of 5 or above indicating cellulitis needing treatment, a score of 3 or 4 needing dermatology consultation, and a score of 2 or lower unlikely to be cellulitis.

Using this method for 175 participants, researchers found 21 (12%) individuals had scores 2 or lower and 84 (48%) had scores of 3 or 4. Of the 70 (40%) with scores of 5 or higher, 40% were found to have a false-positive ALT-70 result and instead have pseudocellulitis.

Both ALT-70 and thermal imaging methods had a high sensitivity score (98.8% and 93.5%, respectively), whereas combining the two methods showed the best specificity at 53.9%.

“Cellulitis results in skin surface temperature elevations significantly greater than in pseudocellulitis,” the authors wrote. “Surface thermal imaging may represent a useful adjunct to the clinical assessment of cellulitis (reducing overdiagnosis), but additional validation of diagnostic performance and cutoff values in more diverse populations is needed.”