Diagnosing pulmonary aspergillosis in patients with hematological malignancies: a multicenter prospective evaluation of an Aspergillus PCR assay and a galactomannan ELISA in bronchoalveolar lavage samples

Abstract

Diagnosing invasive pulmonary aspergillosis (IPA) remains a challenge in patients with hematological malignancies. The clinical significance of testing bronchoalveolar lavage (BAL) samples both with polymerase chain reaction (PCR) and Aspergillus galactomannan ELISA (GM) is unclear, and the BAL cutoff for GM has not been clearly evaluated yet. Using a validated nested PCR assay and a GM ELISA, we prospectively examined BAL samples from 87 hematological patients at high risk of IPA. Of 76 (87%) evaluable patients, 29 patients had proven or probable disease. The receiver operating characteristic (ROC) analysis of GM optical density (OD) cutoff levels yielded a BAL OD of 0.5 to be optimal. We identified 29 probable or proven cases based on this OD. Sensitivity and specificity for BAL GM were 0.79 (95% CI, 0.62-0.9) and 0.96 (95% CI, 0.86-0.99), respectively. For BAL PCR, sensitivity and specificity were 0.59 (95% CI, 0.41-0.75) and 0.87 (95% CI, 0.75-0.94), respectively. Combining BAL GM and PCR for diagnosis showed a sensitivity and specificity rate of 0.55 (95% CI, 0.38-0.72) and 1.0 (95% CI, 0.93-1.0), respectively, if positivity was defined by positive results for both tests. If either positive BAL GM or positive BAL PCR results defined test positivity, the sensitivity was 0.83 (95% CI, 0.65-0.92), and the specificity was 0.83 (95% CI, 0.70-0.91) In terms of optimal sensitivity and specificity, a GM OD cutoff of 0.5 was determined for BAL samples. Positivity for both GM and Aspergillus PCR in BAL makes a pulmonary aspergillosis highly likely.