A serum galactomannan and neutrophil-to-lymphocyte ratio-based nomogram for predicting in-hospital mortality in non-neutropenic invasive pulmonary aspergillosis

作者：Xinyu Wang, Wenjuan Li, Yao Ma, Yi Li, Shengyan Ding, Jiayi Shen, Tingting Zhao, Yajie Lu, Chao Sun, Xin Su · 发表于：Frontiers in Cellular and Infection Microbiology · 年份：2025 · DOI：10.3389/fcimb.2025.1675277 · 被引用次数：3 · 研究领域：Antifungal resistance and susceptibility、Pleural and Pulmonary Diseases、Interstitial Lung Diseases and Idiopathic Pulmonary Fibrosis

Background: We developed a novel prognostic model combining serum galactomannan (GM) and neutrophil-to-lymphocyte ratio (NLR) for predicting in-hospital mortality in non-neutropenic invasive pulmonary aspergillosis (IPA). Methods: We retrospectively identified 278 eligible IPA patients using the automated natural language processing system. Patients were divided into survival (n=199) and non-survival (n=79) groups. A multivariate logistic regression was developed to predict in-hospital mortality based on age, critical condition, serum GM and NLR. Internal validation was performed using bootstrap resampling methods. Subsequently, a clinical nomogram was constructed to support clinical decision-making. Results: Serum GM exhibited a higher specificity of 78.9% (95% CI: 72.6-84.3%), whereas NLR demonstrated relatively greater sensitivity at 70.9% (95% CI: 59.6-80.6%) for mortality prediction. The combination of either positive exhibited a sensitivity of 81.0%, while the dual-positive criterion achieved a specificity of 90.5%. A comprehensive model integrating age, critical condition, serum GM, and NLR demonstrated robust discrimination, with a mean AUC of 0.815 (SD = 0.005) after bootstrap resampling with 1000 iterations. This result was consistent with the original model AUC of 0.818 (95% CI: 0.767-0.870). Decision curve analysis further confirmed that the comprehensive model provided a greater net benefit compared to the simple model based solely on age and critical condition, ...