Background: Proximity extension assay (PEA) is a novel antibody-based proteomic technology. Sparse data have been published concerning the matrix effect of serum vs. ethylenediamine tetraacetic acid (EDTA) plasma and the reproducibility of results obtained using PEA technology. Methods: We analyzed samples with the PEA-based 92-plex Olink® immuno-oncology (I-O) assay. To estimate the matrix effect, we analyzed paired serum and EDTA plasma samples from 12 patients with biliary tract cancer. To evaluate the reproducibility, we used data from 7 studies, where 6–8 serum samples from patients with pancreatic cancer were used as bridging samples on 3 versions of the panel over a 2.5-years period. Results: For the study of serum vs. plasma, 80 proteins were evaluable. The mean serum to EDTA plasma ratio ranged from 0.41–3.01. For 36 proteins, the serum and plasma values were not comparable due to high variability of the ratio, poor correlation, or possible concentration effect. For the bridging samples, the mean intra-study inter-assay coefficient of variation (CV) ranged from 11.3% to 26.1%. The mean inter-study CV was 42.0% before normalization and 26.2% after normalization. Inter-study results were well correlated (r ≥ 0.93), especially for studies using the same version of the panel (r ≥ 0.99). Conclusion: For 44 of 92 proteins included in the Olink® I-O panel, the variation between results obtained using serum and EDTA plasma was constant and results were well correlated. Furthermore, samples could be stored for several years and used on different versions of the same PEA panel without it effecting results.