**4. Personalized medicine and kinase inhibitors**

The basic concept of personalized cancer medicine is to match the patient with an appropriate therapeutic regimen that will provide the best outcome [30]. This can be accomplished on a variety of different levels. With regard to a pharmacological approach, the traditional way patients are assigned to therapy is based on tumor type, that is, what anatomical site at which the tumor resides and/or what type of histopathology the tumor cells derive from. Once these pieces of information were obtained the patient would be assigned a pharmacotherapy regimen that provided the best statistical outcome based on collective experience. This approach has worked as well as can be expected, and exceedingly well in some tumor types considering the diseases' high mortality. However with modern scientific advances and strive toward better outcomes, an even more personalized approach has emerged.

A modern approach to personalized cancer pharmacotherapy can be considered on three basic levels which can be applied individually or combined. These three levels are the patient's pharmacogenetics, the tumor's genomics and/or proteomics, and the tumor's response to drug exposure. Patient pharmacogenetics refers to the patients genetics that are relevant to effects on pharmacokinetics of drugs, for example, polymorphisms in drug metabolizing enzymes. Tumor genomics and/or proteomics refers to specific genetic abnormalities that effect specific gene expression and is thought to contribute to tumor development or maintenance. The application of any or all of these pieces of information may improve the outcome of pharma‐ cotherapy for the patient.

The application of pharmacogenomics to personalized therapy can be exemplified by appli‐ cation to tamoxifen therapy for the breast cancer patient. Tamoxifen acts as an anti-estrogen on breast tissue and is used for treatment of estrogen receptor positive (ER+) breast cancer. It has been determined that tamoxifen is converted in vivo to 4-hydroxytamoxifen, which is a much more active anti-estrogen agent. It was subsequently found that this transformation is accomplished by the drug metabolizing enzymes cytochrome P-450 (CYP) isofroms 2D6 and 2C19 [31]. Patients that were receiving tamoxifen therapy and concurrently receiving selective serotonin reuptake inhibitor (SSRI) drugs, known to inhibit CYP 2D6, experienced poor outcomes because of the failure to convert tamoxifen to 4-hydroxtamoxifen. A significant percentage of the population carry a genetic polymorphism in the CYP 2D6 gene which results in poor metabolism with respect for the CYP 2D6 isoform and will not receive the full benefit of tamoxifen therapy [32]. Consequently a readily available genomic test is available for patients that can identify those who are not good candidates for tamoxifen therapy, because they have a specific 2D6 polymorphism, so that they can be directed to alternatives.

Advances in genomics and proteomics have enabled selection of patients that may benefit from targeted therapies for certain tumor types. Analysis of tumor cells on the protein level using immunohistochemistry (IHC) can identify cells that express relevant protein targets. An example of the application of this approach is the detection of c-erbB2 (Her2/Neu) receptor in breast cancer patients which can direct them toward trastuzumab therapy. Specific gene mutations in tumor cells may be detected using techniques of RT-PCR, DNA microarray, and fluorescence in situ hybridization (FISH) that can be used to direct patients to therapy that may be beneficial. For example, a study demonstrated the utility of screening non-small cell lung cancer patients for amplification of the epidermal growth factor receptor (EGFR). Patients with amplified EGFR receiving an EGFR kinase inhibitor (gefitinib) had longer median progression free survival (PFS) and overall survival (OS), compared to those who did not have amplified EGFR [33, 34]. Furthermore, of those patients who responded to gefitinib therapy 77% had EGFR gene amplification, whereas only 33% of non-responders had EGFR gene amplification. This study demonstrates the utility of identifying patients with mutations that drive tumor growth and survival and matching them with appropriate targeted therapy. The study also highlights the shortcoming of single gene determination directed therapy in that 33% of non-responders also had EGFR amplification.

A potentially more invasive and perhaps technically challenging approach is to directly test samples of the patient's tumor against available drugs. This approach of using personalized xenografts to direct patient therapy was demonstrated in a patient with advanced pancreatic patient [35]. After initial surgery metastases were discovered and adjuvant gemcitabine therapy failed to halt progression. A personalized mouse xenograft model was developed from the patient's tumor tissue and was found to respond to the DNA alkylating agents, mitomycin C and cisplatin. The patient was assigned to mitomycin C treatment and subsequently cisplatin and achieved a partial response with duration of 50+ months. Genomic analysis of the patient's tumor tissue revealed inactivation of the PALB2 gene, which is involved with repair of double strand DNA breaks. It seems logical that the patient's tumor responded to the DNA alkylating agents that would cause double stranded DNA breaks.

The genomic approach has been applied to kinase inhibitors from the beginning of their introduction to the arsenal of pharmacotherapy options. Imatinib, the first kinase inhibitor to be marketed is targeted at a specific genomic alteration, a chromosomal translocation produc‐ ing the Philidelphia chromosome (Ph) that expresses a mutant gene product, the BCR-Abl kinase. This kinase is constitutively active, and is the driver of nearly all chronic myeloid leukemias (CML). Therefore, patients with CML and are Ph+ can be matched to imatinib therapy. Imatinib therapy has been fairly successful for CML patients. It has been shown that patients who achieve complete cytogenic response at 2 years on imatinib therapy tend to maintain the durable response and do not have mortality significantly different than the general population [36].

Because the effectiveness of kinase inhibitors require that the target kinase to be a driver of, expressed in, or aberrantly expressed in, or be mutated in the tumor cells, it is important to know the status of the target within the specific patient's tumor in order to best assign the patient to kinase inhibitor therapy. To this end specific diagnostic tests have been developed to help guide selection of kinase inhibitor therapy. Indeed, twelve of the approved kinase inhibitors' prescribing information assume diagnostic testing be performed, and of those, four require a diagnostic test for prescription. Table 2 lists six such diagnostics recognized by the USFDA [37].
