Immune responses to cancer cells can be elicited in vivo by administering synthetic peptides derived from proteins uniquely or overexpressed on tumor cells (tumor associated antigens - TAAs). Peptides derived from TAAs are presented in the context of major histocompatibility complex (MHC) molecules to cytotoxic T cells (CTL), which can recognize and lyze tumor cells. In contrast to peptides derived from an exogenous source (viral or bacterial), tumor peptides bind weakly to MHC class I molecules. The low binding affinity of these peptides makes them poor candidates for vaccination due to the poor immunogenic response produced. In order to enhance antigen recognition and hence immunogenicity, peptide binding affinity for MHC can be initially improved by modifying the 'anchor' residues. However, the task at hand is highly unpredictable and minor changes in peptide sequence can alter/abolish the T cell response. Furthermore, despite the wealth of information obtained over the last decade from high resolution X-ray structures of MHC class I in complex with peptides (pMHC) as well as pMHC in complex with T cell receptor (TCR), prediction remains difficult. Nonetheless, peptides represent convenient chemical entities that can be rapidly synthesized in clinical grade for therapeutic applications. Herein, the rationale behind modifying TAAs will be discussed including the synthesis/use of proteolytically tolerant peptides (and peptide mimetics) which incorporate non-natural amino acids, retro-inversion and cyclization to improve bioavailability.