Donor-specific HLA antibodies in predicting crossmatch outcome: Comparison of three different laboratory techniques

The virtual crossmatch, which is based on single antigen bead technology, is used in the prediction of crossmatch results. However, this assay differs in sensitivity and specificity from crossmatch methods. In our study, the results of physical crossmatches, performed with three different methods, were assessed against virtual crossmatch results. The aim was to determine the potential cut-off values for donor specific antibodies (DSA) that would predict the crossmatch results obtained by different methods. The results of different crossmatch techniques were correlated with the virtual crossmatch. The receiver operating characteristic (ROC) analysis revealed the Flow cytometric crossmatch (FCXM) and Luminex crossmatch (LXM) to be the most accurate, with area under curve (AUC) values of 0.861 and 0.805, respectively. While we found that the virtual crossmatch correlated well with all the crossmatch results, FCXM produced the best results (83% of the DSA detected). LXM outperformed the other tests in terms of the accuracy in separating class II DSA.