BackgroundStudies on the prognostic value of the soluble programmed death ligand 1 (sPD-L1) in cancer patients have not yielded consistent results.ObjectiveThis meta-analysis was performed to assess the association between sPD-L1 and the prognosis of cancer patients.MethodsPublished articles in Pubmed, EMBASE, and Cochrane clinical trial databases were searched from the inception to September 2020. Overall survival (OS), progression-free survival (PFS), recurrence-free survival (RFS), and disease-free survival (DFS) data were evaluated using a hazard ratio (HR) at 95% confidence interval (95% CI).ResultsA total 31 studies involving 17 tumors and 3,780 patients were included. The overexpression of sPD-L1 was associated with shorter OS (HR 1.85, 95% CI 1.59–2.15, I2 = 33%). High sPD-L1 had worse PFS (HR 2.40, 95% CI 1.55–3.72, I2 = 83%), and worse DFS (HR 2.92, 95% CI 2.02–4.29, I2 = 40%), without significant statistical difference in RFS (HR 2.08, 95% CI 0.99–4.40, I2 = 0%).ConclusionsHigh sPD-L1 levels were associated with worse survival prognosis in cancer patients. The sPD-L1 may be a potential prognostic, non-invasive, and dynamic monitoring biomarker for cancers in the future.