Clinical utility of non-induced sputum galactomannan testing for suspected invasive pulmonary aspergillosis: a real-world retrospective cohort study.

PURPOSE: Non-induced sputum galactomannan (GM) testing is increasingly used in routine evaluation of suspected invasive pulmonary aspergillosis (IPA), yet its real-world performance and the clinical implications of negative results remain incompletely characterised. We aimed to evaluate its diagnostic performance relative to bronchoscopic mycological findings and to describe the short-term evolution of GM results in routine practice.

METHODS

We retrospectively analysed adult patients who underwent non-induced sputum GM testing between 2017 and 2025. Diagnostic performance was assessed in patients with paired sputum and bronchoscopic samples obtained within 7 days using a bronchoscopy-based composite mycological reference standard (defined as bronchoscopic GM index ≥ 0.5, positive Aspergillus culture, or histopathological evidence of invasion). Longitudinal GM evolution was described in patients with initially negative sputum GM results.

RESULTS

Among 1002 patients, 73 had paired samples. Sputum GM demonstrated a sensitivity of 73.7% (95% CI 53.9-93.5%), specificity of 72.2% (95% CI 60.3-84.2%), positive predictive value of 48.3% (95% CI 30.1-66.5%), and negative predictive value of 88.6% (95% CI 79.3-98.0%) for composite bronchoscopic mycological positivity. Higher sputum GM indices were associated with increasing rates of composite bronchoscopic mycological positivity (P for trend = 0.003). Among patients with initially negative sputum GM who underwent repeat testing, early follow-up (days 8-30) showed GM signal evolution in 15.1% of retested patients, whereas subsequent serum or bronchoscopic GM positivity remained uncommon (3.8%); both proportions increased with longer observation.

CONCLUSION

In routine clinical practice, non-induced sputum GM testing demonstrates moderate diagnostic performance relative to bronchoscopic mycological findings. Negative results were associated with low short-term rates of subsequent bronchoscopic or serum GM positivity. These findings may inform short-term clinical reassessment, particularly in settings where bronchoscopy is not readily available.