# Integrative diagnosis of invasive pulmonary aspergillosis in non-neutropenic patients using BALF-tNGS–derived Aspergillus load and host risk factors: a multicenter study

## TL;DR

This study improves the diagnosis of a serious fungal lung infection in patients without neutropenia by combining genetic testing results with patient risk factors.

## Contribution

A new diagnostic model for invasive pulmonary aspergillosis in non-neutropenic patients using tNGS fungal load and host risk factors.

## Key findings

- The model achieved high diagnostic accuracy (AUC = 0.966) for invasive pulmonary aspergillosis.
- Patients with medium Aspergillus load had the highest prevalence of invasive pulmonary aspergillosis.
- Key risk factors included diabetes, corticosteroid use, and ICU admission.

## Abstract

Diagnosing invasive pulmonary aspergillosis (IPA) in non-neutropenic patients is challenging because of non-specific manifestations and limited diagnostic tools. Targeted next-generation sequencing (tNGS) of bronchoalveolar lavage fluid (BALF) enables rapid pathogen detection; however, its capacity for quantitative assessment of fungal load remains unclear. This study integrated BALF-tNGS fungal load with host risk factors to develop a diagnostic nomogram for IPA in non-neutropenic patients.

Non-neutropenic adults with suspected IPA were retrospectively enrolled at three tertiary hospitals (December 2020–December 2024). IPA was classified according to consensus definitions. Normalized Aspergillus reads from BALF-tNGS were stratified into low, medium, and high tiers. Clinical, radiological, and microbiological variables were analyzed, and a multivariable logistic regression model was built and internally validated via bootstrapping. Diagnostic performance was assessed using receiver operating characteristic analysis.

Among 238 patients, 134 had IPA and 104 had non-IPA. Compared with non-IPA cases, patients with IPA had higher rates of diabetes (73.9%), corticosteroid exposure (87.3%), bacterial co-infection (94.8%), and intensive care unit (ICU) admission (72.4%) (all P < 0.01). IPA prevalence peaked in the medium-load group (75.4%) compared with the low-load (36.7%) and high-load (38.3%) groups (P < 0.001). Seven independent predictors, Aspergillus load, diabetes, corticosteroid exposure, bacterial co-infection, ICU admission, nodular shadow, and positive BALF culture, were incorporated into the model, which showed excellent discrimination (AUC = 0.966; sensitivity, 91.8%; specificity, 95.2%) and good calibration.

Integrating BALF-tNGS–derived normalized Aspergillus reads with host factors substantially improves differentiation of IPA from colonization in non-neutropenic patients. This semi-quantitative framework supports early, individualized antifungal decision-making and merits external validation.