Principles of Oncologic Pharmacotherapy

The effective use of cancer chemotherapy requires a thorough understanding of the principles of neoplastic cell growth kinetics, basic pharmacologic mechanisms of drug action, pharmacokinetic and pharmacodynamic variability, and mechanisms of drug resistance. Recent scientific advances in the field of molecular oncology have led to the identification of large numbers of potential targets for novel anticancer therapies. This has resulted in a tremendous expansion of the drug development pipeline, and in the present era, the diversity of clinically useful novel anticancer therapeutic agents is growing at an unprecedented rate. However, the great enthusiasm that surrounds these new agents must be tempered by the challenges they present in optimizing their clinical use and in rationally integrating them with existing anticancer therapies. This discussion focuses on the basic principles underlying the development of modern combination chemotherapy, and it is followed by a description of the major classes of chemotherapeutic drugs and their mechanisms of action.

The cell cycle and tumor growth pattern

The growth pattern of individual neoplastic cells may greatly affect the overall biologic behavior of human tumors and their responses to specific types of cancer therapy. Tumor cells can be subdivided into three general populations: (1) cells that are not dividing and are terminally differentiated; (2) cells that continue to proliferate; and (3) nondividing cells that are currently quiescent but may be recruited into the cell cycle. The kinetic behavior of dividing cells is best described by the concept of the cell cycle.

The cell cycle is composed of four distinct phases during which the cell prepares for and undergoes mitosis. The G₁ phase consists of cells that have recently completed division and are committed to continued proliferation. After a variable period, these cells begin to synthesize DNA, marking the beginning of the S phase. After DNA synthesis is complete, the end of the S phase is followed by the premitotic rest interval called the G₂ phase. Finally, chromosome condensation occurs and the cells divide during the mitotic M phase. Resting diploid cells that are not actively dividing are described as being in the G₀ phase. The transition between cell-cycle phases is strictly regulated by specific signaling proteins; however, these cell-cycle checkpoints may become aberrant in some tumor types.

Some anticancer agents induce their cytotoxic effects during specific phases of the cell cycle. Antimetabolites, such as fluorouracil (5-FU) and methotrexate, are more active against S-phase cells, whereas the vinca alkaloids, epipodophyllotoxins, and taxanes are relatively more M-phase specific. These kinetic properties may have clinically important consequences for cancer chemotherapy. For example, cell-cycle–nonspecific agents, such as the alkylating agents and platinum derivatives, generally have linear dose-response curves (ie, increasing the dose increases cytotoxicity). In contrast, cell-cycle–specific agents will often plateau in their concentration-dependent effects because only a subset of proliferating cells remain fully sensitive to drug-induced cytotoxicity. These cell-cycle–specific agents tend to be schedule dependent, because the only way to increase the total cell kill is by extending the duration of exposure, not by increasing the dose.

In clinical practice, solid tumors typically have low growth fractions and heterogeneous doubling times; as they increase in size, tumors may outgrow their blood and nutrient supply, leading to slower growth rates. In real life, most tumors display a sigmoid-shaped Gompertzian growth pattern in which growth rates decline as tumors expand. The most rapid growth occurs at small tumor volumes, whereas larger tumors may harbor higher numbers of nonproliferating cells, potentially making them less sensitive to agents that selectively target dividing cells. This understanding of tumor growth kinetics has been used to support the development of novel clinical strategies for optimizing cancer chemotherapy. This includes the use of adjuvant chemotherapy to treat small volumes of tumor cells during times of high growth rates and the sequential administration of non–cross-resistant drug combinations.

Principles of combination chemotherapy

Based upon cell kinetic and pharmacologic principles, a set of guidelines for designing modern combination chemotherapy regimens has been derived. Multiagent therapy has three important theoretical advantages over single-agent therapy. First, it can maximize cell kill while minimizing host toxicities by using agents with nonoverlapping dose-limiting toxicities. Second, it may increase the range of drug activity against tumor cells with endogenous resistance to specific types of therapy. Finally, it may also prevent or slow the development of newly resistant tumor cells. Specific principles for selecting agents for use in combination chemotherapy regimens are listed in Table 1.

Definition of response

In clinical studies, formal response criteria have been developed and have gained wide acceptance. The National Cancer Institute (NCI) has proposed and implemented newer standard response criteria called Response Evaluation Criteria in Solid Tumors (RECIST). In contrast, the World Health Organization (WHO) has a different standard for assessing response. Major differences between these guidelines are outlined in Appendix 1.

Drug resistance

Drug resistance to chemotherapy may arise from a variety of different mechanisms, including anatomic, pharmacologic, and genetic processes. Some of the common factors that may broadly affect tumor cell sensitivity to different classes of agents include the failure of drugs to penetrate into specific sanctuary sites, such as the brain and testes, or the development of mutations in the target proteins that render them less sensitive to specific molecular inhibitors. Another factor may be decreased drug accumulation resulting from the increased expression of drug efflux pumps in the cell membrane, such as p-glycoprotein, which is encoded for by the multidrug resistance (MDR-1) gene. This 170-kDa glycoprotein normally removes toxins or xenobiotic metabolites from cells via an energy-dependent process. High levels of MDR-1 expression in tumor cells correlate with resistance to a wide range of cytotoxic agents. Other drug efflux pumps that have been implicated in the development of broad resistance to cancer chemotherapy include the MDR-associated protein (MRP) and breast cancer resistance protein (BCRP).

Pharmacokinetic and pharmacodynamic variability

Variability in clearance

The rational clinical use of cancer chemotherapy requires a thorough understanding of the variability in human response to drug therapy. One of the major goals of the field of clinical pharmacology is to precisely define the processes responsible for this variability in drug action. Pharmacokinetic variability can arise from interindividual differences in drug adsorption, distribution, metabolism, and excretion. All of these processes result in differences in drug delivery to its ultimate site of action. In contrast, pharmacodynamic variability arises from inherent differences in the sensitivity of target tissues to drug effects. Both kinetic and dynamic factors can complicate the treatment of individual cancer patients and must be addressed by the practicing oncologist on a daily basis. Although a formal review of drug pharmacokinetics and pharmacodynamics is not possible here, a brief discussion of the most clinically relevant points is warranted.

The most clinically useful parameter in drug therapy is clearance, because clearance reflects all the processes in the body that contribute to drug elimination. In oncology, the importance of clearance is enhanced because clearance is the only parameter that relates dose to the measured area under the concentration vs time curve (AUC), which is a useful measure of systemic drug exposure. Mathematically, clearance is defined as dose/AUC. Clearance is not a rate of drug elimination; instead it is defined as a volume of drug cleared per unit of time.

Overcoming interindividual variation in clearance is a fundamental goal of pharmacokinetic analyses. Because they tend to be highly toxic with low efficacy, anticancer drugs may have the narrowest therapeutic index of any class of agents used in clinical medicine. Thus, the ability to administer an individualized dose of drug to achieve a uniform target AUC and a uniform clinical result is often a high priority for cancer therapeutics. Because clearance defines the relationship between dose and AUC, estimating clearance prior to anticancer drug administration is extremely important. A common attempt to individualize cancer chemotherapy is to dose a drug based upon the body surface area (BSA) expressed in mg/m² to achieve a uniform AUC in patients with different body sizes. Inherent in this approach is the fundamental assumption that clearance is strongly correlated with BSA. However, when studied formally, the relationship between clearance and BSA is often weak and does not consistently justify the routine use of this dosing approach. Nonetheless, although the application of BSA-based dosing has been widely criticized, it still remains a common practice.

Recognizing clinical situations in which drug clearance is commonly altered, such as in patients with hepatic or renal dysfunction, is important for agents that are eliminated by these routes. For example, carboplatin is extensively cleared by glomerular filtration, and its systemic AUC in plasma ultrafiltrates is strongly correlated with pharmacodynamic effects, such as thrombocytopenia. Thus, dosing strategies that estimate the glomerular filtration rate (GFR) to achieve a targeted AUC and thereby minimize excessive toxicity in individual patients have gained wide clinical acceptance. Likewise, for hepatically metabolized drugs, doses must be adjusted in patients with liver dysfunction. However, accurately assessing hepatic drug-metabolizing capacity is more difficult than estimating the GFR. Nonetheless, guidelines for dose-adjusting agents metabolized in the liver, such as doxorubicin, have been established.

Pharmacogenomic variability

Variability in drug action may also be caused by genetic factors. The new field of pharmacogenomics attempts to define the impact of genetic differences on drug kinetics and dynamics. McLeod and Evans have defined pharmacogenomics as the field of study of “the inherited nature of interindividual differences in drug disposition and effects.” Clinically relevant genetic variations have been characterized in specific drug-metabolizing enzymes, such as the cytochrome P-450 isoforms CYP2D6 and CYP2C9.

In the field of medical oncology, well-defined examples of clinically relevant pharmacogenetic differences are relatively few; however, this is an important area of ongoing research. Perhaps the best characterized example in clinical oncology is the inherited deficiency of the thiopurine methyltransferase (TPMT) enzyme that results in severe intolerance to thiopurine therapy. Another example is the inherited variation in dihydropyrimidine dehydrogenase (DPD) activity, the rate-limiting catabolic enzyme responsible for clearance of 5-FU. Genetic alterations associated with DPD deficiency have been identified in rare patients experiencing severe and fatal toxicity after treatment with standard doses of 5-FU. Finally, interindividual variation in irinotecan-induced toxicities may be partially explained by genetic polymorphisms in the gene encoding the UDP-glucuronosyltransferase (UGT) 1A1 enzyme that is involved in the clearance of the active metabolite SN-38. Also, the identification of somatic mutations in the tyrosine kinase domain of the epidermal growth factor receptor (EGFR) in non–small-cell lung cancer (NSCLC) and their correlation with response to EGFR inhibitors has become an important event in the fields of cancer genomics and therapeutics. The issue of ethnic diversity in the pathogenesis of given tumors was raised by the initial observation of a higher response to gefitinib (Iressa) and erlotinib (Tarceva) in patients of Asian origin, together with the discovery that they harbor more frequent EGFR mutations in non–small-cell lung cancer.

Pharmacodynamic polymorphisms that directly affect target tissue sensitivity to drug effects are also clinically important. For example, polymorphisms in the promoter region of the thymidylate synthase gene have been correlated with tumor response to 5-FU–based chemotherapy that targets this enzyme. In the near future, our understanding of how genetics affects a drug’s pharmacokinetics may ultimately allow for the optimization of specific treatment regimens for individual patients. Likewise, our understanding of how genetics affects pharmacodynamic variation may be enhanced by powerful new technologies that can characterize the expression of literally thousands of genes within the tumor itself. The molecular profiling of tumor cells by DNA microarray techniques and other advances in biotechnology offer tremendous hope for improving our ability to treat cancer in the near future.

Molecularly targeted therapies

At the beginning of the 21st century, important and meaningful advances in anticancer therapeutics are being discovered at an unparalleled rate. Much of this progress is driven by the explosion of knowledge in the field of molecular oncology. The sequencing of the entire human genome has dramatically increased the number of promising molecular targets for new and novel anticancer treatment strategies. These advances hold great promise for developing a new generation of agents with high specificity for tumor cells (see discussion later in this chapter).

Chemotherapeutic agents classified by mechanism of action

Alkylating agents

The alkylating agents impair cell function by forming covalent bonds with the amino, carboxyl, sulfhydryl, and phosphate groups in biologically important molecules. The most important sites of alkylation are DNA, RNA, and proteins. The electron-rich nitrogen at the 7 position of guanine in DNA is par­ticularly susceptible to alkylation.

Alkylating agents depend on cell proliferation for activity but are not cell-cycle­phase–specific. A fixed percentage of cells are killed at a given dose. Tumor resistance probably occurs through efficient glutathione conjugation or by enhanced DNA repair mechanisms. Alkylating agents are classified according to their chemical structure and mechanism of covalent bonding; this drug class includes the nitrogen mustards, nitrosoureas, and platinum complexes, among other agents (Table 2).

Nitrogen mustards The nitrogen mustards, which include such drugs as mechlorethamine (Mustargen), cyclophosphamide, ifosfamide, and chlorambucil (Leukeran), are powerful local vesicants; as such, they can cause problems ranging from local tissue necrosis to pulmonary fibrosis to hemorrhagic cystitis. The metabolites of these compounds are highly reactive in aqueous solution, in which an active alkylating moiety, the ethylene immonium ion, binds to DNA. The hematopoietic system is especially suscep­tible to these compounds.

Nitrosoureas The nitrosoureas are distinguished by their high lipid solubility and chemical instability. These agents rapidly and spontaneously decompose into two highly reactive intermediates: chloroethyl diazohydroxide and isocyanate. The lipophilic nature of the nitrosoureas enables free passage across membranes; therefore, they rapidly penetrate the blood-brain barrier, achieving effective CNS concentrations. Consequently, these agents are used for a variety of brain tumors.

Platinum agents Cisplatin is an inorganic heavy metal complex that has activity typical of a cell-cycle-phase–nonspecific alkylating agent. The compound produces intrastrand and interstrand DNA cross-links and forms DNA adducts, thereby inhibiting the synthesis of DNA, RNA, and proteins. Carboplatin has the same active diamine platinum moiety as cisplatin, but it is bonded to an organic carboxylate group that allows increased water solubility and slower hydrolysis to the alkylating aqueous platinum complex, thus altering toxicity profiles. Oxaliplatin (Eloxatin) is distinguished from the other platinum compounds by a di-amino-cyclohexane ring bound to the platinum molecule, which interferes with resistance mechanisms to the drug.

Antimetabolites

Antimetabolites are structural analogs of the naturally occurring metabolites involved in DNA and RNA synthesis. As the constituents of these metabolic pathways have been elucidated, a large number of structurally similar drugs that alter the critical pathways of nucleotide synthesis have been developed.

Antimetabolites exert their cytotoxic activity either by competing with normal metabolites for the catalytic or regulatory site of a key enzyme or by substituting for a metabolite that is normally incorporated into DNA and RNA. Because of this mechanism of action, antimetabolites are most active when cells are in the S phase and have little effect on cells in the G₀ phase. Consequently, these drugs are most effective against tumors that have a high growth fraction.

Antimetabolites have a nonlinear dose-response curve, such that after a certain dose, no more cells are killed despite increasing doses (5-FU is an exception). The antimetabolites can be divided into folate analogs, purine analogs, adenosine analogs, pyrimidine analogs, and substituted urea.

Natural products

A wide variety of compounds possessing antitumor activity have been isolated from natural substances, such as plants, fungi, and bacteria. Likewise, selected compounds have semisynthetic and synthetic designs based on the active chemical structure of the parent compounds, and they, too, have cytotoxic effects.

Antitumor antibiotics Bleomycin preferentially intercalates DNA at guanine-cytosine and guanine-thymine sequences, resulting in spontaneous oxidation and formation of free oxygen radicals that cause strand breakage.

Anthracyclines The anthracycline antibiotics are products of the fungus Streptomyces percetus var caesius. They are chemically similar, with a basic anthracycline struc­ture containing a glycoside bound to an amino sugar, daunosamine. The anthracyclines have several modes of action. Most notable are intercalation between DNA base pairs and inhibition of DNA–topoisomerases I and II. Oxygen free radical formation from reduced doxorubicin intermediates is thought to be a mechanism associated with cardiotoxicity.

Epipodophyllotoxins Etoposide is a semi­synthetic epipodo­phyllotoxin extracted from the root of Podophyllum peltatum (man­drake). It inhibits topoisomerase II activity by stabilizing the DNA–topoisomerase II complex; this process ultimately results in the in­ability to synthesize DNA, and the cell cycle is stopped in the G1 phase.

Vinca alkaloids The vinca alkaloids are derived from the periwinkle plant Vinca rosea. Upon entering the cell, vinca alkaloids bind rapidly to the tubulin. The binding occurs in the S phase at a site different from that associated with paclitaxel and colchicine. Thus, polymeriza­tion of microtubules is blocked, resulting in impaired mitotic spindle formation in the M phase.

Taxanes Paclitaxel and docetaxel (Taxotere) are semisynthetic derivatives of extracted precursors from the needles of yew plants. These drugs have a novel 14-member ring, the taxane. Unlike the vinca alkaloids, which cause microtubular disassembly, the taxanes promote microtubular as­sembly and stability, therefore blocking the cell cycle in mitosis. Docetaxel is more potent than paclitaxel in enhancing microtubular assembly and also induces apoptosis.

Camptothecin analogs include irinotecan and topotecan (Hycamtin). These semisynthetic analogs of the alkaloid camptothecin, derived from the Chinese ornamental tree Camptotheca acuminata, inhibit topoisomerase I and interrupt the elongation phase of DNA replication.

Targeted agents

Monoclonal antibodies Although monoclonal antibodies (MAbs) have been used in cancer therapeutics since the late 1990s, the number of new agents in this class is growing exponentially. Several unconjugated MAbs have established utility in medical oncology as highly targeted therapies. The earliest therapeutic MAb to show convincing utility in medical oncology was rituximab (Rituxan), approved in 1997 for the treatment of non-Hodgkin lymphoma. This antibody targets the CD20 antigen found on B-cell lymphocytes and can be used clinically as a single agent or in association with combination chemotherapy. Another MAb, trastuzumab (Herceptin), has shown excellent activity in combination with chemotherapy in breast cancer patients whose tumor cells overexpress the human epidermal growth factor receptor 2 (HER2) protein. Finally, alemtuzumab (Campath) is a MAb that recognizes the CD52 antigen expressed on both B-cell and T-cell lymphocytes. This agent is useful in the treatment of chemotherapy-refractory B-cell chronic lymphocytic leukemia.

In 2004, two new MAbs were approved by the US Food and Drug Administration (FDA) for the treatment of patients with advanced colorectal cancer: bevacizumab (Avastin) and cetuximab (Erbitux). Bevacizumab binds to the vascular endothelial growth factor (VEGF) and prevents ligand-induced VEGF receptor activation, which blocks the stimulation of endothelial cell growth and inhibits new blood vessel formation in tumors that secrete VEGF. Cetuximab binds the EGFR on the surface of tumor cells, ultimately leading to down-regulation of this signaling pathway. This process blocks tumor growth and proliferation and can reverse tumor resistance to chemotherapeutic agents such as irinotecan. Cetuximab is now also indicated for head and neck cancer. Bevacizumab has received indications in lung and breast cancers.

Panitumumab (Vectibix), approved by the FDA in 2006, is another MAb that targets EGFR and is indicated for use in colorectal cancer that has metastasized following standard chemotherapy.

A recent randomized trial found that the addition of bevacizumab to irinotecan- and 5-FU–based chemotherapy in newly diagnosed patients with advanced colorectal cancer significantly improved the response rate (45% vs 35%; P = .0029), duration of response (10.4 vs 7.1 months; P = .0014), and median survival (20.3 vs 15.6 months; P = .0003) compared with placebo. Grade 3 hypertension (10.9% vs 2.3%) and GI perforations (1.5% vs 0%) were more common in the bevacizumab arm, but overall, the bevacizumab therapy was thought to be well tolerated. This was the first randomized clinical trial demonstrating a survival benefit for antiangiogenic therapy.

Small-molecule targeted therapies Molecularly targeted therapies are designed to selectively interact with specific molecular pathways within cells to achieve a rational antitumor effect. The classic rationally designed molecularly targeted agent is imatinib (Gleevec), which was identified in screening studies designed to detect inhibitors of the Bcr-Abl tyrosine kinase, present in virtually all cases of chronic myelogenous leukemia. Originally synthesized as an inhibitor of platelet-derived growth factor receptor (PDGFR), it is also a potent inhibitor of the c-kit tyrosine kinase. Imatinib binds to the ATP-binding site and inhibits the tyrosine kinase’s ability to phosphorylate its substrates.

Gefitinib (Iressa) is another small-molecule–targeted therapy that is a signal transduction inhibitor of the EGFR tyrosine kinase. It binds noncovalently to the ATP-binding site of the ­intracellular domain of the EGFR protein and blocks the kinase activity. Its anticancer effects arise from the ability to interfere with EGFR-mediated signaling, which is associated with cell proliferation, angiogenesis, and cell motility.

Erlotinib (Tarceva) is an HER1/EGFR-targeted therapy that has demonstrated a significant survival benefit as second-line therapy for patients with advanced NSCLC. A randomized phase III study performed by the NCI of Canada compared erlotinib with best supportive care in patients with recurrent or refractory non–small-cell lung cancer following treatment with either one or two lines of prior chemotherapy. As reported by Shepherd et al, erlotinib was significantly better than best supportive care in terms of response rate (8.9% vs < 1%; P < .001), disease progression-free survival (2.23 vs 1.84 months; P < .001), and median survival (6.7 vs 4.7 months; P = .001). Rash and diarrhea were the most frequent side effects, although only 5% of patients discontinued erlotinib due to drug-related toxicities. This is the first randomized trial of an EGFR inhibitor in lung cancer demonstrating a significant survival benefit.

Another targeted therapy is the 26S proteasome inhibitor bortezomib (Velcade). The 26S proteasome is a ubiquitous multiprotein complex responsible for degrading a variety of regulatory proteins involved in cancer cell proliferation. In multiple myeloma cells, bortezomib induces apoptosis by mechanisms that are not precisely defined. One hypothesis is that the inhibition of the proteasomal degradation of I-κB, an inhibitor of the transcription factor Nuclear Factor kappa B (NF-κB), prevents the constitutive activation of NF-κB. In multiple myeloma cells, NF-κB is thought to be necessary for cell proliferation and survival.

Sunitinib (Sutent) is an oral small molecule multitargeted tyrosine kinase inhibitor with antiproliferative and antiangiogenic activity attributed to the inhibition of PDGFR, VEGFR, KIT, and FLT₃. In a phase II study by Motzer et al, involving 63 patients with metastatic renal cell carcinoma, sunitinib resulted in a 33% partial response rate (95% confidence interval [CI], 22% to 46%), a 37% rate of stable disease lasting at least 3 months, and a 1-year median survival of 65% (95% CI, 50% to 76%). Treatments were well tolerated, with major grade 3 or 4 toxicities consisting of lymphopenia, neutropenia, elevated amylase/lipase levels without associated pancreatitis, and fatigue or asthenia.

Sorafenib (Nexavar) is a new oral small multitargeted molecule with antiproliferative and antiangiogenic activity attributed to the inhibition of PDGFR, VEGFR, KIT, FLT³, and, RAF-1.

Temsirolimus (Torisel) is a specific inhibitor of mTOR, a signaling protein that regulates cell growth and angiogenesis. It has been evaluated in a phase III, randomized, three-arm study versus interferon-alpha or the combination of both agents in the treatment of first-line, poor-risk patients with advanced renal cell carcinoma. Results have shown this agent to be the first to produce a significant increase in overall survival with an acceptable safety profile.

Sunitinib, sorefenib, and temsirolimus all show significant activity in advanced renal cell cancer patients, frequently producing prolonged stabilization of disease, and in the case of sunitinib, objective remissions. These targeted therapies have displaced cytokines as the gold standard for treating advanced-stage renal cell tumors.

Other newer molecularly targeted therapies include dasatinib (Sprycel), approved by the FDA for treatment of chronic myelogenous leukemia and Philadelphia-positive acute lymphoblastic leukemia, and lapatinib (Tykerb), approved by the FDA for treatment of advanced or metastatic Her2-positive breast cancer in combination with capecitabine (Xeloda).

The relevance of pharmacogenomics to molecularly targeted therapies was highlighted by the impressive finding that specific somatic mutations localizing to the ATP-binding site of the EGFR tyrosine kinase in lung tumors were associated with clinical response. In eight of nine patients with NSCLC who responded to gefitinib therapy, specific mutations were identified, whereas none was found in seven matched nonresponders. This landmark finding, coupled with what is already known about c-kit mutations that “drive” the proliferation of gastrointestinal stromal cell tumors in response to imatinib therapy, suggests that it may ultimately be possible to prospectively select patients with NSCLC who have a high probability of responding to EGFR tyrosine kinase inhibitors. Further studies are under way to confirm these findings and to extend the observation to other tumor types.

See Appendix 3 for a list of drugs and/or new indications recently approved for treatment of cancer and Appendix 4 for the uses, dosages, and toxicities of the chemotherapeutic agents discussed in this chapter.

Suggested reading

Cunningham D, Humblet Y, Siena S, et al: Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 351:337–345, 2004.

Hudes G, Carducci M, Tomczak P, et al: Temsirolimus, interferon alfa, or both for advanced renal cell cancer. N Engl J Med 356:2271–2281, 2007.

Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335–2342, 2004.

Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350:2129–2139, 2004.

McLeod HL, Evans WE: Pharmacogenomics: Unlocking the human genome for better drug therapy. Annu Rev Pharmacol Toxicol 41:101–121, 2001.

Morabito A, De Maio E, Di Maio M, et al: Tyrosine kinase inhibitors of vascular endo- thelial growth factor receptors in clinical trials: Current status and future directions. Oncologist 11:753–764, 2006.

Motzer RJ, Hutson TE, Tomczak P, et al: Sunitinib versus interferon alfa in metastatic renal-cell carcinoma. N Engl J Med 356:115–134, 2007.

Olopade OI, Schwartsmann G, Saijo N, et al: Disparities in cancer care: A worldwide perspective and roadmap for change. J Clin Oncol 24:2135–2136, 2006.

Paez JG, Janne PA, Lee JC, et al: EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy. Science 304:1497–1500, 2004.

Shepherd FA, Pereira J, Ciuleanu TE, et al: Erlotinib in previously treated non–small-cell lung cancer. N Engl J Med 353:123–132, 2005.

Weinshilboum R: Inheritance and drug response. N Engl J Med 348:529–537, 2003.