Aspergillus galactomannan detection in the diagnosis of invasive aspergillosis in cancer patients

Abstract

Purpose: To assess the Aspergillus galactomannan enzyme-linked immunosorbent assay (ELISA) in the diagnosis of invasive aspergillosis (IA) in adult and pediatric oncohematologic patients.

Patients and methods: The study was conducted in four patient groups: those with fever of unknown origin (FUO) during neutropenia, suspected pulmonary infection (PI), or nonpulmonary aspergillosis (NPA) and those undergoing surveillance (S) after hematopoietic stem-cell transplantation (HSCT). IA was classified as definite, probable, or possible, according to European Organization for Research and Treatment of Cancer/Mycosis Study Group definitions.

Results: A total of 3,294 serum samples were collected during 797 episodes (FUO, 261; PI, 297; NPA, 28; and surveillance, 211), and 153 episodes of IA were diagnosed (31 definite, 67 probable, and 55 possible). Three episodes were first suspected from galactomannan ELISA; the remaining 150 cases were diagnosed from clinical or radiologic evidence. Sensitivity of the ELISA was 64.5%, 16.4%, and 25.5% in definite, probable, and possible episodes of IA, respectively, and was lower in patients positive for anti-Aspergillus antibodies than in antibody-negative patients. Most false-positive results occurred in children and in allogeneic HSCT (allo-HSCT) patients. Overall specificity of the ELISA was 94.8%. It was lower in children compared with adults (P <.0001) and in allo-HSCT patients compared with non-allo-HSCT adults (P =.0002). Lowering the ELISA cutoff value from 1.500 to 0.700 seemed more relevant for non-allo-HSCT adults (sensitivity, 73.1%, 44.3%, and 44.7% in definite, probable, and possible IA, respectively; specificity, 94%).

Conclusion: Galactomannan ELISA seems less sensitive than previously described, and sensitivity can be further reduced by the presence of anti-Aspergillus antibodies. A new cutoff value for the ELISA of 0.700 is proposed for non-allo-HSCT adults.