Aim To evaluate cytokine and soluble programmed death ligand-1 (sPD-L1) levels in the serum and plasma of cancer patients treated with immunotherapy, and to test different assays. Methods Three Luminex xMAP assays and two ELLA microfluidic cartridges were used to screen 28 immune-related biomarkers in 38 paired serum and citrate-theophylline-adenosine-dipyridamole (CTAD) plasma samples collected from 10 advanced melanoma or non-small cell lung cancer (NSCLC) patients at different time points during immunotherapy. Results Twenty-three of 28 biomarkers were detected both in serum and plasma by at least one of the assays, including IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, GM-CSF, IFN-γ, TNF-α, VEGF, IP-10, MCP-1, eotaxin, fractalkine, G-CSF, IFN-α, IL-1RA, IL-13, IL-17A, MIP-1β and sPD-L1. Conversely, FGF-2 and IL-1α were not detected in both matrices; GRO-α factor and EGF were detected only in serum and MIP-1α only in plasma. sPD-L1, MCP-1, IFN-γ, IL-8, MIP-1β and VEGF were, respectively, 1.15-, 1.44-, 1.83-, 2.43-, 2.82-, 6.72-fold higher in serum, whereas IL-10, IL-4, IL-2 and IL-5 were 1.05-, 1.19-, 1.92- and 2.17-fold higher, respectively, in plasma. IP-10 levels were higher in plasma but, as well as for VEGF, the bias serum versus plasma varied depending on the assay used (IP-10: −5.7% to −145%; VEGF: 115% to 165%). No significant differences were found for the remaining nine analyzed cytokines. Conclusion The cytokine and sPD-L1 levels may differ between serum and plasma samples collected from cancer patients treated with immunotherapy, and the results obtained can be influenced by the different characteristics of the tested assays. The standardization of pre-analytical and analytical procedures is therefore needed for the future implementation of these circulating biomarkers in clinical practice.