Abstract
BACKGROUND: We aimed to assess whether soluble programmed death-ligand 1 (sPD-L1) could serve as a new biomarker for PTB. METHODS: Plasma sPD-L1 levels in the discovery cohort were analyzed through flow cytometry and validated by sandwich ELISA. Pleural effusion sPD-L1 levels were measured using ELISA. RESULTS: In the discovery cohort, sPD-L1 levels in the severe (SE, n = 44), non-severe (non-SE, n = 34) and HC (n = 10) group were 67.41 (30.14-126.41), 26.75 (11.00-52.35) and 14.6 (10.78-21.91) pg/ml, respectively. The sPD-L1 levels in SE patients were significantly higher than those in both non-SE patients and HCs (p < 0.0001). These findings were confirmed in the validation cohort with sPD-L1 levels significantly higher in SE (n = 60,763.81 pg/ml) compared to both non-SE patients (n = 80, 318.30 pg/ml) and HCs (n = 79, 202.33 pg/ml)(p < 0.0001). Receiver operating characteristic (ROC) analysis demonstrated plasma sPD-L1 could distinguish SE from non-SE PTB with an AUC of 0.8058 (95% CI 0.7308-0.8808). sPD-L1 levels showed positive correlations with inflammatory markers, such as neutrophil percentage (NEU%, r = 0.5743, p < 0.0001), neutrophil-to-lymphocyte ratio (NLR, r = 0.5952, p < 0.0001). Survival analysis revealed shorter survival times in groups with higher sPD-L1 (≥445.1 pg/ml, p = 0.0006). In addition, sPD-L1 levels in tuberculous pleural effusion (TPE) were significantly higher than malignant pleural effusion (MPE) (1964.72 versus 159.38 pg/ml, p < 0.001), showing diagnostic performance (AUC = 0.9837) similar to adenosine deaminase (AUC= 0.9859). CONCLUSION: Elevated plasma sPD-L1 may be a predictive marker for both disease severity and poor prognosis in PTB patients. Pleural effusion sPD-L1 levels might potentially function as an adjunctive marker for differentiating TPE from MPE.