# A Prospective Observational Study on the Evaluation of the Galactomannan Test in the Diagnosis of Invasive Aspergillosis

## TL;DR

This study evaluates the effectiveness of the galactomannan test in diagnosing invasive aspergillosis, finding it more accurate than traditional methods when using a higher threshold.

## Contribution

The study provides new evidence that BAL GM testing at OD > 1.0 outperforms conventional diagnostic methods for invasive aspergillosis.

## Key findings

- BAL GM testing at OD ≥ 0.5 showed 93.3% sensitivity and 71.4% specificity for diagnosing IA.
- BAL GM at OD > 1.0 improved specificity to 88.6% and accuracy to 89.9%, outperforming KOH and HRCT.
- Serum GM had lower sensitivity (84.0%) and specificity (80.0%) compared to BAL GM.

## Abstract

Background: Invasive aspergillosis (IA) is a life-threatening opportunistic fungal infection, particularly in immunocompromised patients. Early, reliable diagnosis is essential for timely management.

Objective: To assess the diagnostic performance of the galactomannan enzyme immunoassay (GEIA) in bronchoalveolar lavage (BAL) fluid and serum, comparing with potassium hydroxide (KOH) microscopy, high-resolution CT (HRCT) chest, and culture (as reference).

Methods: In this prospective, single-center observational study (January-December 2018), BAL and/or serum samples from patients clinically suspected of IA were tested using the Platelia™ Aspergillus Galactomannan EIA kit (Bio-Rad, France). Results at two optical density (OD) cut‑offs (≥ 0.5 and > 1.0) were compared against culture, KOH microscopy, and HRCT. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and 95% confidence intervals were calculated. Pairwise comparisons used McNemar’s test.

Results: Among 159 BAL samples, GEIA positivity was 51.6% (OD ≥ 0.5) and 39.0% (OD > 1.0). At OD ≥ 0.5, BAL galactomannan (GM) sensitivity was 93.3% (95% CI: 83.8-97.9), specificity 71.4% (95% CI: 61.4-79.9), PPV 63.4%, NPV 97.4%, and accuracy 79.9%. At OD > 1.0, sensitivity was 92.6% (95% CI: 82.9-97.9), specificity 88.6% (95% CI: 80.8-93.8), PPV 80.7%, NPV 95.9%, and accuracy 89.9%. KOH microscopy and HRCT in the BAL cohort achieved sensitivities of 90.7% and 85.2% and specificities of 77.1% and 80.0%, respectively. McNemar’s test showed BAL GM (OD > 1.0) was significantly superior to KOH (p = 0.003) and HRCT (p = 0.005). In 60 patients tested with serum GM (OD ≥ 0.5), sensitivity was 84.0% (95% CI: 62.9-95.0), specificity 80.0% (95% CI: 61.4-92.3), PPV 75.0%, NPV 87.5%, and accuracy 81.7%.

Conclusions: In this cohort, BAL GM testing provided excellent sensitivity and, especially at OD > 1.0, improved specificity relative to conventional methods. Serum GM performed moderately well, but with lower sensitivity and specificity. BAL GM (OD > 1.0) may provide an optimal balance for non-culture-based diagnosis of IA, though results should be interpreted in the clinical context. Future multicenter studies incorporating polymerase chain reaction (PCR) and β-D-glucan are warranted to validate these findings.