Thermal imaging combined with a predictive model that relies on routine clinical and laboratory parameters proved better at ruling out cellulitis in the emergency department (ED) compared with either test on its own, and could potentially reduce overdiagnosis of the common bacterial skin infection, a prospective validation study found.
Alone, thermal imaging to measure the maximum skin surface temperature of the affected leg and the ALT-70 cellulitis prediction model -- which incorporates limb asymmetry, leukocytosis, tachycardia, and age 70 or older -- yielded high sensitivity (93.5% and 98.8%, respectively) but lower specificity (38.4% and 22%).
But with the two methods combined, sensitivity remained high, at 91.7%, while specificity improved to 53.9%, reported researchers led by Michael Pulia, MD, PhD, of the University of Wisconsin-Madison School of Medicine and Public Health, in .
"Assessment of skin temperature by touch is a core component of the physical examination when evaluating patients with potential cellulitis," Pulia told 51˶ in an email. "The thermal imaging technology takes this subjective assessment of skin temperature and makes it objective."
In the study of more than 200 ED patients, the team found that thermal imaging of the maximum skin surface temperature of the affected leg in patients with a consensus cellulitis diagnosis was significantly higher than in those with conditions that mimic cellulitis (33.2°C vs 31.2°C, P<0.001).
"This study represents the best evidence yet supporting thermal imaging as a useful adjunct to clinical judgement in the diagnosis of cellulitis," Pulia said.
The combined approach used a thermal cutpoint of 31.2°C or higher and a score of 3 or higher on the ALT-70 prediction model to indicate cellulitis -- scores of 5 or above on the 0-7 scoring system indicate cellulitis requiring treatment with oral antibiotics or intravenous antibiotics in serious cases.
Overdiagnosis of cellulitis is an important issue because other conditions are mistaken for cellulitis in up to one-third of cases, wrote Arash Mostaghimi, MD, MPH, of Brigham and Women's Hospital in Boston, and David Li, MD, MBA, of Boston Dermatology and Laser Center, in an .
Misdiagnosis of cellulitis has been estimated to result in 50,000 to 130,000 avoidable hospitalizations every year. Eliminating misdiagnoses could result in a potential savings of as much as $515 million per year and avoid tens of thousands of unnecessary courses of antibiotics and hospital admissions annually, they commented.
The accuracy of both and the in diagnosing cellulitis have been evaluated individually in previous studies, Mostaghimi and Li wrote, and noted that the current study "reflects the latest step in a quest for improved diagnosis and treatment of cellulitis." However, they pointed out that there are still no diagnostic markers for cellulitis that could be used for "a quick and economic diagnosis at the point of care, especially in outpatient and urgent care settings," where rapid consultation with a dermatologist is not always feasible.
Both modalities alone or in combination had high negative predictive values -- 95.2% for ALT-70, 87.8% for maximum skin surface temperature, and 87.5% for both in combination. Positive predictive values were 53.9%, 55.5%, and 64.7%, respectively.
The study included 204 participants who presented to the ED with acute dermatologic lower extremity symptoms and visible erythema from October 2018 to March 2020. Of the participants, 45% had a diagnosis of cellulitis confirmed by a consensus panel of six board-certified physicians, while the rest were determined to have pseudocellulitis. The mean age of study participants was 56.6 years and 59% were men. Most participants were white (88%), which was a study limitation along with the single-center design.
Patients were excluded if they had high-risk or complicated wounds, or symptoms that only involved the toes. During the ED visit, photographs and skin surface temperature images were taken of both the participants' affected and unaffected limbs using the commercially available FLIR ONE Pro thermal camera. Software analyzed skin surface temperatures and generated data on maximum and mean temperatures.
Approximately 20% of participants with confirmed cellulitis had a fever while in the ED versus 6.3% of those with pseudocellulitis. Venous insufficiency and stasis dermatitis were the most common diagnoses among participants with pseudocellulitis, the researchers reported.
Disclosures
The study was funded by the Agency for Healthcare Research and Quality and the National Center for Advancing Translational Sciences.
Pulia and co-authors disclosed no relationships with industry.
Mostaghimi disclosed relationships with Hims & Hers, AbbVie, Sun Pharmaceuticals, Pfizer, Digital Diagnostics, Lilly, Equillium, ASLAN, Boehringer Ingelheim, ACOM, and Olaplex. Li disclosed no relationships with industry.
Primary Source
JAMA Dermatology
Pulia MS, et al "Validation of thermal imaging and the ALT-70 prediction model to differentiate cellulitis from pseudocellulitis" JAMA Dermatol 2024; DOI: 10.1001/jamadermatol.2024.0091.
Secondary Source
JAMA Dermatology
Li DG, Mostaghimi A "The quest for a more accurate diagnosis of cellulitis: In dreams begin responsibilities" JAMA Dermatol 2024; DOI: 10.1001/jamadermatol.2024.0089.