Abstract
Automatic data processing of radioimmunoassay results is now one of the most important parts in the radioimmunoassay laboratories. Especially in handling all batches of samples, it allows us to process radioimmunoassay results with simplicity, rapidity and reproducibility. Recently, some instrumentation appeared which could count and calculate the sample concentration automatically using some type of approximation of the standard curves. Rectangular hyperbola is employed for IRI, a-FP, HGH etc., but approximation of HTSH standard curves using rectangular hyperbola is impossible.
Theoretically, linearization of the standard curves of HTSH are possible using logit transformation of bound percent (Y) and log concentration (X) in the case of univalent homogeneous antigen-antibody reaction. However, this is not always true in the routine radioimmunoassays because of several factors associated with the procedure.
From these reasons, authors divided the standard curve into two parts at the point of middle concentration (31.2μU/ml) and calculated each linear regression line from logit (Y) and loge (X) of the standard curve (two-divided logit transformation method) . These calculations were made using a small desk-top electronic program calculator (Canola 1614P) . After determination of coefficients of each regression line, sample HTSH values were calculated semiautomatically using newly developed calculation program.
In the 113 samples correlation coefficient between two-divided method and usual graphic method was highly significant (γ=0.9977, p<0.001) . To compare this, correlation coefficient between single logit transformation and graphic method was γ=0.9906 (p <0.001), and correlation coefficient of our method slightly improved compare to the single logit transformation.
Almost all the standard curves obtained by plotting loge (X) against logit (Y) showed single monophasic curves, and the inflexion point lay near at the middle point of the standard curves. This may be variable according to the change in the composition of the radioimmunoassay kits, but relatively stable at the middle point of the concentration. This may be the reason why our method well correlated with usual graphic method.
Our method employes only a small desk-top electronic calculator and simple handling manner, and could be done semiautomatically. And it was found to be useful in clinical bases, even in the small laboratories.
