<p><strong>OBJECTIVES: </strong>Cutaneous lesions are often the first marker of invasive mold infection, which can cause substantial morbidity in immunocompromised children. The purpose of this study was to describe the evaluation and outcomes of immunocompromised children who presented with findings requiring skin biopsy because of concern about invasive infection. In children who were biopsied, we sought to determine the factors predictive of invasive mold infection.</p>

<p><strong>METHODS: </strong>A retrospective review was conducted at the Children's Hospital of Philadelphia. Patients included in the study were immunocompromised individuals younger than 26 years old who underwent skin biopsy by the inpatient dermatology consultation team between January 1, 2003, and March 15, 2015, because of development of new cutaneous lesions that were suspected of being invasive infection.</p>

<p><strong>RESULTS: </strong>One hundred five encounters met the inclusion criteria. Fifty (47.6%) biopsied individuals had an infectious pathogen identified on histopathology or culture. Mold was the most common (36%) pathogen, followed by bacteria (32%) and yeast (26%). The presence of a single lesion (P = .001) and prior occlusion at the site of the lesion (P &lt; .001) were associated with mold on biopsy. The combination of a single lesion, history of occlusion, and tissue necrosis on examination was highly predictive for invasive mold infection (86.3% [95% confidence interval 55.1-97.0%]). Of the 18 individuals with confirmed invasive mold infection, 13 (72%) underwent surgical resection, of whom 12 (92%) survived the 30-day follow-up period.</p>

<p><strong>CONCLUSION: </strong>Skin biopsy enabled the detection of a pathogen that informed directed therapeutic interventions in nearly half of participants. Institutions caring for immunocompromised children should ensure adequate staffing of clinical personnel approved to perform skin biopsies.</p>

<p>Frontline clinicians caring for hospitalized children typically knew the indication for antimicrobial therapy but less often knew the current day or planned duration of therapy or of plans for intravenous to oral conversion. Night shift clinicians were less likely to know day of therapy and duration of therapy than day shift clinicians caring for the same patients.</p>