BACKGROUND Invasive pulmonary aspergillosis (IPA) is a major cause of morbidity and mortality in patients with hematological malignancies in the setting of profound neutropenia and/or hematopoietic stem cell transplantation. Early diagnosis and therapy has been shown to improve outcomes, but reaching a definitive diagnosis quickly can be problematic. Recently, galactomannan testing of bronchoalveolar lavage (BAL) fluid has been investigated as a diagnostic test for IPA, but widespread experience and consensus on optical density (OD) cut-offs remain lacking. METHODS We performed a prospective case-control study to determine an optimal BAL galactomannan OD cutoff for IPA in at-risk patients with hematological diagnoses. Cases were subjects with hematological diagnoses who met established definitions for proven or probable IPA. There were two control groups: subjects with hematological diagnoses who did not meet definitions for proven or probable IPA and subjects with non-hematological diagnoses who had no evidence of aspergillosis. Following bronchoscopy and BAL, galactomannan testing was performed using the Platelia Aspergillus seroassay in accordance with the manufacturer's instructions. RESULTS There were 10 cases and 52 controls. Cases had higher BAL fluid galactomannan OD indices (median 4.1, range 1.1-7.7) compared with controls (median 0.3, range 0.1-1.1). ROC analysis demonstrated an optimum OD index cutoff of 1.1, with high specificity (98.1%) and sensitivity (100%) for diagnosing IPA. CONCLUSIONS Our results also support BAL galactomannan testing as a reasonably safe test with higher sensitivity compared to serum galactomannan testing in at-risk patients with hematological diseases. A higher OD cutoff is necessary to avoid over-diagnosis of IPA, and a standardized method of collection should be established before results can be compared between centers.