Background: As microbiological documentation of invasive aspergillosis (IA) is difficult to obtain, non
culture-based methods, such as measurement of fungal biomarkers by serological methods and
detection of Aspergillus DNA by PCR, are important adjunctive diagnostic tools. Data concerning the
performance of these assays for surveillance of IA in high-risk patients in Greece are scarce. We
therefore conducted a multicentre study evaluating the performance of the detection of current
surrogate biomarkers, galactomannan (GM) and 1,3-β-D glucan (BDG), and an Aspergillus real-time
PCR assay in serum samples from high-risk hematological patients.
Material/methods: A total of 104 patients with hematological malignancies (58 AML, 12 ALL, 8 MDS,
5 NHL and 31 other) were screened between 4/2013-7/2015 in four hospitals in the area of Athens,
Greece. Serial blood specimens were collected twice-weekly and tested for GM (Platelia Aspergillus
EIA), BDG (Fungitell) and Aspergillus DNA detection (White J.Mol.Diagn. 2006). The number of
evaluable serum samples was 271. For most of patients there was one sample before neutropenia
and several samples (3-15) during neutropenia. Patient episodes were stratified according to
EORTC/MSG 2008 consensus criteria. The three assays commonly used in clinical microbiology
laboratories were compared head to head and collusions about their performance were made. Positive
GM index, which is included as a microbiological criterion in the diagnosis of probable IA, was used as
a surrogate marker for calculating sensitivity/specificity rates and positive/negative predictive values
(PPV/NPV) of BDG and PCR alone and in combination.
Results: The agreement was 64% between GM-BGD, 77% between GM-PCR and 59% between BGD-PCR
assays. Of 271 samples, 130 (48%) were positive in at least one biomarker whereas only 10 (4%)
were positive in all three assays. Of 104 patients, 63 (61%) were positive in at least one biomarker
whereas only 8 (8%) were positive in all three assays. The test characteristics for each sample, i.e.
sensitivity, specificity, PPV and NPV, PCR and assays were 74%, 63%, 20%, and 95% for BGD, 68%,
78%, 27%, 95% for PCR, and 82%, 45%, 15% and 96% for both BGD/PCR, respectively.
Conclusions: Among the three biomarkers, higher agreement was found between BDG and GM. The
BDG and PCR had good sensitivity and specificity (63-78%) and excellent negative predictive value
(95%) which was not increased significantly when the two assays were combined.
Full conference title:
- ECCMID 27th (2017)