In order to protect from tumour progression, tumour antigen specific CD8 T cells need to be activated in a robust manner (Pardoll, 2003). Extended analysis of biological effects of immunotherapy revealed that tumour specific human T cell responses were often insufficient, making it impossible to draw firm conclusions on the role of T cells in clinical responses. Even in animal models, it remains difficult to identify correlates of T cell mediated protection from tumour progression. In infectious diseases, several T cell criteria have been established as correlates of protection. First, the capacity of T cells to efficiently interact with cognate antigen (Alexander-Miller et al., 1996; Bennett et al., 2007; Derby et al., 2001; Gallimore et al., 1998; Messaoudi et al., 2002; Sedlik et al., 2000; Speiser et al., 1992; Yee et al., 1999; Zeh et al., 1999), a property which is often termed as ‘functional avidity’ and which is primarily controlled by T cell receptors (TCRs). However, a survey of oligo or monoclonal CD8 T cell responses to MAGE-A3 vaccination in melanoma patients who had objective tumour regression failed to show high avidity of antigen recognition (Connerotte et al., 2008). Second, a central role is played by the precursor frequency of specific T cells (Moon et al., 2007; Obar et al., 2008). Finally, protection depends on the functional capacity of T cells, likely dependent on functional avidity and many more parameters. Recent attempts have focused on direct ex vivo functional profiling of T cells, in order to identify multiple functional properties of effective T cells that are operational in vivo (Almeida et al., 2007; Betts et al., 2006; Daucher et al., 2008; Lichterfeld et al., 2007; Rehr et al., 2008). There is increasing consensus that T cells must be multifunctional (Appay et al., 2008; Seder et al., 2008). Mouse models suggest that for tumour immunity, similar principles may apply as for infectious diseases. Thus, functional avidity appears to be essential, including the capacity of T cells to recognize naturally expressed and processed tumour antigen (Kochenderfer and Gress, 2007). High avidity CD4 T cells would also be critical in determining tumour rejection mediated by CD8 T cells (Brandmaier et al., 2009). Precursor frequency of appropriate T cells plays also a key role (Rizzuto et al., 2009). Finally, it is important that T cells are fully functional, with competent responsiveness to antigen, proliferation, T cell survival, homing, effector function and generation of immunological memory (Appay et al., 2008; Klebanoff et al., 2006; Wherry et al., 2003). In the adoptive transfer setting, it is increasingly clear that T cells with the ability to persist for prolonged periods of time in vivo correlate with clinical efficacy (Berger et al., 2008; Heemskerk et al., 2008; Shen et al., 2007; Zhou et al., 2005). Improvement of immunotherapy may depend on careful optimization towards the generation of T cells fulfilling these criteria. Results from small scale phase I/II clinical studies provide the rational basis to select the most promising treatment modalities for subsequent costly phase III trials assessing clinical efficacy.