(A–D) Proportions of patients with bronchial obstruction (BO), necrosis, bronchiectasis, and mortality in the thrombosis versus nonthrombosis groups (n = 1,508). Categorical variables were compared using χ² tests and are presented as percentages. (E and F) Comparison of hospitalization duration (E) and costs (F) between thrombosis and nonthrombosis groups (n = 1,508). Data are presented as median (IQR) and compared using the Mann-Whitney U test. (G) LASSO coefficient profiles of candidate laboratory indicators. Each curve represents the trajectory of a variable’s regression coefficient as a function of log (λ) (n = 1508). The vertical dashed line indicates the optimal penalty parameter (λ) selected by cross-validation. (H) Forest plot of multivariable Firth’s penalized logistic regression analysis in the propensity score–matched (PSM) cohort (n = 303). Neutrophil count, C-reactive protein (CRP), D-dimer, and preadmission fever duration were independently associated with thrombosis. Odds ratios (ORs) with 95% CIs and corresponding P values are shown. (I) Comparison of peripheral neutrophil counts between patients with and without thrombosis. Violin plots illustrate the distribution of neutrophil levels in each group (n = 303). Statistical significance was assessed using the Mann-Whitney U test. (J) Receiver operating characteristic (ROC) curve evaluating the discriminative performance of neutrophil count for thrombosis (n = 303). Internal validation was performed using bootstrap resampling (1,000 iterations). The bootstrap-corrected area under the curve (AUC) was 0.899, with a 95% CI of 0.802–0.957. (K) Kaplan-Meier curves showing the cumulative probability of thrombosis stratified by neutrophil count (high versus low) in the PSM cohort (n = 303). Differences between groups were assessed using the log-rank test. The number of patients at risk at each time point is shown below the plot. ***P < 0.001, ****P < 0.0001.