Performance of diagnostic algorithms in patients with invasive pulmonary aspergillosis.

Abstract

Background: Invasive pulmonary aspergillosis (IPA), once limited to immunocompromised patients, is now a severe complication in critically ill ICU patients without classic risk factors. Due to the difficulty of obtaining histological evidence, its diagnosis relies on poorly tested algorithms in real-world settings.

Methods: We conducted a retrospective multicenter (n=9) cohort study including 202 patients with IPA. Patients were classified using a multistep process based on the European Organization for the Research and Treatment of Cancer/Mycosis Study Group (EORTC-MSG), Invasive-Fungal Diseases in Adult Patients in Intensive Care Unit (FUNDICU), Aspergillus ICU (Asp-ICU), and Asp-ICU with biomarkers (Asp-ICU-BM) criteria. We then evaluated the predictive performance of these criteria against the clinical cohort and histologically proven cases.

Results: Among 202 patients, 78 had EORTC-MSG host factors and were classified accordingly, with the EORTC-MSG criteria achieving 100% agreement in identifying clinical and histologically proven cases. In 112 ICU patients without EORTC-MSG host factors, overall agreement was 53% for FUNDICU, 4% for Asp-ICU, and 26% for Asp-ICU-BM compared to the clinical cohort. Validation against histologically proven cases showed FUNDICU had 44% sensitivity and 75% specificity, Asp-ICU 6% sensitivity and 100% specificity, and Asp-ICU-BM 28% sensitivity and 63% specificity. Adding acute respiratory distress syndrome (ARDS) and post-cardiac surgery to the FUNDICU criteria improved sensitivity to 97% with a specificity of 63%. The remaining 12 patients lacked EORTC-MSG host factors and were not in the ICU, highlighting a novel classification system.

Conclusion: EORTC-MSG and FUNDICU IPA classification systems are useful for the assignment of most patients with IPA. Incorporating post-operative complications after cardiac surgery and ARDS enhanced the diagnostic accuracy of FUNDICU.