Résumé : The present research focused on the development of a new methodology to assess the strength of the interaction between vaccine antigens and elicited polyclonal antibodies through SPR biosensors. Quantifying the binding strength of polyclonal antibodies is of first importance to evaluate the quality of the vaccine as well as to increase the scientific knowledge of immune protection mechanisms. To now the development of such tool has been complicated by the non-specific binding caused by high protein abundance in the blood and serum samples but also by the way of interpreting the data resulting from multi-interaction events measured at the same time. At first, we unsuccessfully tried to segregate the individual affinity contribution of each antibody population by measuring the signal as the sum of singular interactions. Differentiation of the singular contribution would have needed the fulfillment of the “additivity” hypothesis, meaning that each antibody bind identically alone or in mixture with other antibody. This hypothesis was not met and mathematical assessment by the sum of singular contribution led to fitting results that did not reflect the biological reality. It was therefore decided to switch the analysis method and to measure the end association binding level reached by the different samples injected at the same specific antibody content. The dissociation behavior was interpreted by the percentage of binding after long and fixed dissociation time. In a first application, we compared the antibodies elicited by two different commercially available vaccines and we showed that the binding interaction was not concentration dependent as, highly different levels were reached when injecting identical antibody concentration. No statistical significant difference was observed between both vaccines. Research firstly focused on the decrease of the non-specific binding and we found that ionic strength was a key parameter, increasing the buffer salt concentration reduced the non-specific binding without diminishing the binding strength. The sample composition was also a key parameter and purifying the IgG allowed to decrease dramatically the undesired binding events. A second application aimed at showing the equivalence between two different antigen constructions for two antibodies population. Even if identical antigen level immobilization is a challenge, the methodology is completely suitable to perform a 2-dimensional comparison (ligand and analyte). A last application was dedicated to the comparison between D and Q-pan Flu vaccines, and results showed that there was no statistical evidence of significant differences between both vaccines. End association level correlated well with haemagglutination inhibition assay at least when serum samples were not diluted at the same antibody content. This last application also showed that throughput may be extended to more than 50 samples per 80 hours