Use of regression analysis and the complement-dependent cytotoxicity typing assay for predicting lymphoid chimerism.

The complement-dependent cytotoxicity typing assay was studied for its accuracy in determining the presence of donor lymphocytes within standard chimeric donor/host cell combinations. Regression analysis of the data was utilized to evaluate the chimera assay. Excellent coefficients of determination (r2 greater than 0.90) were obtained for all standard curves, and a significant (P less than 0.001) linear relationship was established between percent cytotoxicity (dependent variable) and level of donor target cell chimerism (independent variable) for each regression equation. A highly significant (P less than 0.001) linear function was also established between percent cytotoxicity and concentration of donor target bone marrow cells. Regression coefficients (slopes) approached, but did not show complete unity (range; b = 0.86-0.95). Therefore, levels of cytotoxicity were not directly equivalent to levels of donor cell chimerism. A more accurate assessment of donor lymphoid chimerism would be provided by regression analysis of standard donor/host cell independent variables and inverse prediction. Significant estimates of peripheral donor lymphoid chimerism in putative mixed chimeric recipients were successfully made by this technique.