Key points are not available for this paper at this time.
Background Squamous cell carcinoma antigen is used as a tumour marker and is routinely measured in clinical laboratories. We validated two different immunoassays and found three cases in which the squamous cell carcinoma antigen concentrations deviated greatly between the two immunoassays. Here, we aimed to elucidate the mechanisms responsible for these deviations. Methods The squamous cell carcinoma antigen concentrations were determined using the ARCHITECT SCC (CLIA method) and the ST AIA-PACK SCC (FEIA method). We performed polyethylene glycol precipitation and size exclusion chromatography to assess the molecular weight and spike recovery and absorption tests to examine the presence of an autoantibody. Results Both methods exhibited good performances for the measurement of squamous cell carcinoma antigen, although a correlation test showed large differences in the squamous cell carcinoma antigen concentrations measured using the two methods in three cases. The results of polyethylene glycol treatment and size exclusion chromatography indicated the existence of a large molecular weight squamous cell carcinoma antigen in these three cases. The spike recovery tests suggested the possible presence of an autoantibody against squamous cell carcinoma antigen. Moreover, the absorption test revealed that large squamous cell carcinoma antigen complexes were formed by the association of squamous cell carcinoma antigen with IgG in two cases and with both IgG and IgA in one case. Conclusions This study describes the existence of large molecular weight squamous cell carcinoma antigen that has complexed with immunoglobulin in the serum samples. The reason for the deviations between the two immunoassays might be due to differences of their reactivities against the squamous cell carcinoma antigen immune complexes with their autoantibody. To our knowledge, this is the first report to describe the coupling of squamous cell carcinoma antigen with IgA.
Mori et al. (Tue,) studied this question.