In the attempt to define a strategy for screening new monoclonal antibodies (mAb) that could be appropriate for clinical application in oncology, we evaluated the suitability of three methods: a direct internalization assay (DIA), an indirect internalization assay (IIA) and an indirect cytotoxicity assay (ICA), by applying them to already selected mAb. The latter were directed against three antigenic systems [38-kDa glycoprotein (gp38), epidermal growth factor receptor, and the neu oncogene product], which, according to their tumor selectivity, could be considered suitable for mAb-guided therapy. The dose-dependent and time-dependent binding, as well as the low intra-assay variability, demonstrated the reliability of the three tests. However, a certain degree of inter-assay variability was observed in each one, the highest value being that found when IIA was applied. Furthermore, the degree of variability, as well as the predictability, seemed to be more related to the mAb/antigen (Ag) combination used rather than to the test applied. From the overall data we suggest a procedure to be applied for screening purposes. As a first approach applied to the raw material, ICA is only suitable for screening in the case of an already selected toxin whereas IIA may be helpful to eliminate the true negative mAb. After purification of the relevant mAb a repeated analysis using DIA could allow the selection of true internalizing mAb. However, this second screening should be followed by a further analysis of the fate of the Ag-Ab complex after internalization.