Detection of serum galactomannan (GM) antigen and presence of the halo sign on a pulmonary computerized tomographic (CT) scan have a high specificity but a low sensitivity to diagnose invasive aspergillosis (IA) in patients at risk for this disease. To our knowledge, the relationship between the time at which pulmonary infiltrates are detected by CT and the time at which GM antigens are detected by enzyme immunoassay (EIA) has not been studied. In a prospective study, tests for detection of GM were performed twice weekly for patients with hematological malignancies who had undergone hematopoetic stem cell transplantation (HSCT) or had received induction and/or consolidation chemotherapy. A pulmonary CT scan was performed once weekly. Infiltrates were defined as either major or minor signs. IA was classified as proven, probable, or possible, in accordance with the definition stated by the European Organization for Research and Treatment of Cancer-Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group. We analyzed 161 episodes of infection in 107 patients (65 allogeneic HSCT recipients, 30 autologous HSCT recipients, and 66 induction and/or consolidation chemotherapy recipients). A total of 109 episodes with no IA, 32 episodes with possible IA, and 20 episodes with probable or proven IA were identified. Minor pulmonary signs were detected by CT in 70 episodes (43%), and major pulmonary signs were detected by CT in 11 episodes (7%). Univariate and multivariate analyses revealed no significant association between detection of GM by EIA and detection of abnormal pulmonary signs by CT. A significant association was found between GM levels and receipt of piperacillin-tazobactam. GM test results were not positive before major signs were seen on CT images. Only 7 (10%) of 70 patients with minor pulmonary signs had positive GM test results before detection of the greatest pathologic change by CT. We show that detection of GM by EIA does not precede detection of major lesions by pulmonary CT. In the clinical setting, the decision to administer mold-active treatment should based on detection of new pulmonary infiltrates on CT performed early during infection, rather than on results of EIA for detection of GM.
