Background: S-100B determination has been shown to be clinically useful in the management of melanoma patients. After the development of a test for determination of the isoforms S100-A1B and S100-BB in serum (S-100B), several sensitive assays for the detection of serum S-100B have become available. We compared four S-100B assays, two automated (LIAISON Sangtec 100 and Elecsys S100) and two manual ones (Sangtec 100 ELISA and CanAg S100 EIA), with respect to clinical data, reference values and correlation.
Methods: In a total of 280 samples from 155 melanoma patients and 98 healthy individuals S-100B values were measured simultaneously with the different assays.
Results: The inter and intra coefficients of variation were best for the automated assays. The functional sensitivity of both manual assays was 0.15 microg/L. Method comparison revealed satisfactory correlation coefficients of 0.9 or higher, but the slopes ranged from 0.29 to 3.36. Except for the Sangtec 100 ELISA, the linearity between the assays was acceptable. The overall sensitivity for melanoma ranged from 37% (Elecsys S100) to 47% (LIAISON Sangtec 100) and the sensitivity increased with stage. ROC curves showed the best accuracy for the LIAISON Sangtec 100 assay.
Conclusions: All assays gave satisfactory results, but it is advisable to improve the performance of the manual assays for better sensitivity. Agreement about an international reference standard is needed.