The Precision Medicine Approach in Oncology

Introduction and Background

Precision Medicine (PM), also known as Personalized or Stratified Medicine, is a broad term used to describe application of the pharmacogenomic and biomarker sciences, along with novel dosing technologies, to the selection of the most appropriate therapeutic agents for individual patients, the monitoring of treatment outcomes and side effects (ADRs) on an individual basis, and the discovery and development of new therapeutic agents with improved efficacy. The overall goal of this approach is to capitalize on the latest developments in technologies relating to genome sequencing and biomarker assays to allow patients to be treated individually by considering their unique genotype and phenotype, thus ensuring optimal efficacy of a therapeutic agent while minimizing side effects. The PM approach has grown rapidly in the last decade, and most pharmaceutical companies have now adopted this overall philosophy or are in the process of moving toward it.

Oncology is the most advanced of the therapeutic areas to have developed and applied PM approaches, partly due to the clinical need to ensure that anticancer therapies are efficacious in individual patients (thus ensuring that time is not wasted) and that side effects are minimized (ensuring comfort and reduced risk for patients). As our understanding of the genetics of cancer has progressed, advances have been made in predicting the risk of disease development and recurrence, in prognosis and staging, in treatment selection and prediction of side effects, and in the facilitation of the discovery and clinical trials of new agents. However, although beyond the scope of this book, progress is now being made in the application of a PM approach in other therapeutic areas, including CNS, cardiovascular, and metabolic diseases (e.g., diabetes).

A tumor, from any given individual, will normally contain a growing number of genetic alterations as it progresses and begins to display an increasing number of characteristics relating to the “Hallmarks of Cancer” (see Chapter 1). Early genetic changes can be used to predict whether patients may be more at risk for developing advanced disease, and whether they might respond to particular therapies. Furthermore, the presence of certain mutated genes in otherwise healthy individuals can be used to predict the risk of cancer occurring. Therefore, genomics (the study of complex sets of genes) and proteomics (the study of their expression and function) are becoming increasingly important. New genetic services are being developed and offered that provide clinically validated individual tumor profiles at the nucleic acid or protein level. In the future, such information may significantly improve the quality of treatment decisions for cancer patients (see Figure 11.1).

The astonishing speed at which science and technology in these areas are progressing means that it must be taught at an appropriate level in the undergraduate and postgraduate curricula of relevant health care professionals such as doctors and pharmacists.

Uses of Genomic Information

In general medicine, much of the current clinical interest in PM is in the variation of genes involved in drug metabolism which can affect both efficacy and side effect profiles, thus influencing drug safety. Therefore, the introduction of PM strategies is viewed as an opportunity to improve both prescribing efficacy and safety. However, there is now growing interest in applying PM approaches to the entire therapeutic process ranging from its use in drug discovery through to patient selection and stratified clinical trials, with an overall goal of ensuring that patients are treated with drugs that work rather than the “one size fits all” approach which was predominant in the latter part of the twentieth century.

An important use of genetic differences in cancer-related genes is in the prediction of risk of disease. During the sequencing of the human genome in the 1990s, powerful genotyping technologies were developed that allowed common variations between individuals to be identified that can be associated with the risk of developing a wide variety of diseases. These variations can be as small as a change to a single nucleotide base within a gene. When a variation is present to a significant extent within a population (e.g., >1%), then it is known as a Single-Nucleotide Polymorphism (abbreviated to SNP, and pronounced “snip”). Interestingly, studies have demonstrated that common genetic variations have only a limited role in determining predisposition to many ordinary diseases, whereas gene variants that are very rare in the general population can have substantial effects on predisposition. For example, rare mutations that cause loss of parts of the genome can increase the risk of CNS diseases such as epilepsy, autism, or schizophrenia by up to twentyfold. SNPs, or more profound variations, or mutations in gene sequences, when associated with increased risk of disease are known as “predisposition disease biomarkers”. As risk factors such as these genetic variations are identified and validated, it is inevitable that there will be a move toward embryonic and

Potential roles of genomics and a “Precision Medicine” approach to cancer risk prediction and treatment (From Genomic Health, Inc.)

FIGURE 11.1 Potential roles of genomics and a “Precision Medicine” approach to cancer risk prediction and treatment (From Genomic Health, Inc.).

other screening programs. Significant progress has already been made in identifying genetic biomarkers for schizophrenia, epilepsy, and some cancers. Although society may embrace screening technologies that identify mutations leading to serious health conditions such as cancer, there is likely to be more debate about the ethics of screening embryos for predisposition to less serious diseases, for example a twentyfold increased risk of a psychiatric disorder.

Another use of these technological advances is to develop “personalized” therapeutic agents tailored to these genetic differences (or biomarkers). A good example of this is the kinase inhibitor imatinib (Gleevec™) used to treat a type of leukemia known as Chronic Myeloid Leukemia (CML). It was initially discovered that CML is caused by a translocation mutation in which parts of tw'o different chromosomes fuse together, leading to a uniquely mutated kinase protein (BCR- ABL) which is only found in the tumor cells of these patients and not their healthy cells. This created an opportunity (exploited by the pharmaceutical company Novartis) to discover a drug that could fit into the ATP-binding pocket of the BCR-ABL protein thus providing a very effective and highly selective treatment (see Chapter 6). Therefore, a pharma- cogenomic assay is carried out on leukemia patients to confirm that they have the BCR-ABL mutation before imatinib is prescribed. Based on this, there is a growing trend for clinicians to order a predictive pre-prescription genetic test to identify potential responders (or nonresponders), or those who are likely to suffer adverse drug reactions (ADRs), before treatment is initiated. Table 11.1 shows a few examples of genes associated with different cancer types.

TABLE 11.1

Examples of Genes Associated with Various Tumor Types.

Cancer Type

Associated Gene(s)


BRCA1. BRCA2. ATM. Her2/neu

Burkitt’s lymphoma


Colon (bowel)



EGFR (ErbB-l). HER2/neu (ErbB-2), HER3 (ErbB-3), HER4 (ErbB-4)

Chronic myelogenous leukemia


Malignant melanoma

CDKN2, BCL-2. B-RAF-V600

Endothelial cancers


Although at present it is possible to identify the up-regulation of single genes that may be associated with more aggressive cancers as shown in Table 11.1), in the future it should be possible to study sets of key relevant genes and the way they interact which should be a more reliable way to inform prognosis and treatment decisions. The way in which genes interact, and the consequences of this for both normal and cancer cells, has given rise to new areas of research such as systems biology and the study of gene networks which involve the use of computational technologies collectively known as bioinfomatics.

To use genomics in cancer diagnostics, prognostics, and treatment, it is necessary to determine which sets of genes and gene interactions are important in different types and subsets of cancers. Analyses can be carried out to link the pattern of gene expression in tumor cells to the individual’s response to therapy or the likelihood of the recurrence of cancer. Results from studies of this type can be used to create a genomic profile (or signature) of an individual’s tumor that, in the future, should allow clinicians to better predict the most beneficial treatments and gain insight into how the cancer is likely to develop. This has led to products such as MammaPrint™ which evaluate multiple genes in cancer patients and help guide therapy (see Section 11.5). However, caution is required when screening for the presence of either single or multiple gene biomarkers (at either the nucleic acid or protein level) in solid tumor biopsy material, due to the heterogeneity of many tumors.

A third use of PM technology is to search a patient’s genome for SNPs relating to individual metabolizing enzymes. This can help predict the response of a patient to a given dose of a drug (e.g., rapid or slow metabolism), or the likelihood of ADRs occurring. For example, individuals with a homozygous polymorphism (i.e., carrying both of the two gene copies) in their UGT1A1 gene (i.e., the “*28 variant”) are unable to metabolize the anticancer agent irinotecan as efficiently as those with a normal version of the gene. Thus, the agent accumulates in these individuals at dose levels considered to be normal in those with nonmutated UGT1 Al. This can lead to significant bone marrow suppression (i.e., neutropenia), which can be fatal. Therefore, in 2005, the FDA made changes to the labeling of irinotecan to indicate that a pharmacogenetic test should be carried out before irinotecan is prescribed, and that patients with the UGT1 Al polymorphism should be given a reduced dose (Table 11.2).

A fourth application of the PM approach is the analysis of the chemical “fingerprints” of metabolic processes in body fluids such as urine, blood, or saliva. This analytical methodology, known as “metabolomics” (or “metabonomics”), attempts to match specific metabolic profiles to particular diseases, or to predict the likelihood of an individual responding

TABLE 11.2

Examples of Genomic Biomarkers Published by the FDA for Inclusion on Drug Labels and the Relevant Test Recommended (1 =Test Required, 2 = Test Recommended, 3 = Information Only) (Adapted from: Chem. Biol. Drug Des., 2007; 69: 381-394, p.383)


Representative Label



Other Drugs Associated with Biomarker



Overexpression of Her2/neu necessary for selection of patients appropriate for drug therapy (breast cancer)




None at present

EGFR expression

Epidermal growth factor receptor presence or absence (colorectal cancer SCCHN: squamous cell carcinoma of head and neck NSCLC, pancreatic cancer)








UGT1A1 variants

UGT1A1 mutation makes patients more susceptible to myelosuppression (colon-rectum cancer)




None at present

TPMT variants

Increased risk of myelotoxicity associated with thiopurine methyltransferase deficiency or lower activity




None at present

С-Kit expression

Gastrointestinal stromal tumor c-Kit expression


Imatinib mesylate (Glivec™)

None at present

Dihydropyrimidine Dehydrogenase (DPD) Deficiency

Severe toxicity (stomatitis, diarrhea, neutropenia and neurotoxicity) associated with deficiency of DPD





Philadelphia Chromosome Deficiency

Philadelphia (Phi) chromosome presence, Busulfan is less efficacious in CML patients lacking the Phi chromosome



None at present

PML/RAR-alpha gene expression (Retinoic Acid Receptor responder and nonresponders)

PML/RAR-alpha fusion gene presence required for drug activity.






Arsenic oxide

to a specific therapeutic agent and/or suffering an adverse drug reaction (ADR) from it. Furthermore, human metabolism is heavily influenced by interactions between the activities of gut microbes and host genes, as well as by diet and environmental factors. This metabolic interplay can directly influence susceptibility to disease, and is being extensively studied. For example, autoimmune disorders, obesity, diabetes, ulcerative colitis, and Crohn’s disease have all been linked to poor gut health and microbial imbalances. Screening using metabolomic approaches, perhaps in association with proteomic or genomic analysis and imaging, is increasingly being used to facilitate early detection of cancer. If identified, then surgery and/or targeted therapies can be initiated with extensive use of prognostic and predictive biomarkers as part of a PM approach.

Many more screens and assays of this type can be expected to be introduced in the future. Most view this as scientific progress, and an opportunity to tailor therapies to individual patients to enhance efficacy and reduce side effects. However, more cynical observers believe that the advantages to patients will be minimal, and that pharmaceutical companies will benefit by marketing costly screening kits alongside their drugs. The pharmaceutical companies usually respond to this assertion by pointing out that this personalized approach to treatment will lead to reduced sales of individual agents as the medical community shifts away from the one-size-fits-all approach that has operated since the dawn of modern therapeutics.

As the technologies to sequence human genomes become faster and less expensive, it is likely that an individual’s genomic sequence will become a key component of their medical record at birth, along with identification of any predisposition disease biomarkers. This should allow the individual to avoid other risk factors for those diseases (e.g., environmental, lifestyle) and to receive appropriate treatment if necessary. For example, lifestyle advice and chemopreventive agents (see Chapter 12) could be introduced to minimize risk. However, there are many ethical boundaries to cross before this becomes a reality. However, some observers are concerned that knowledge of a heightened risk of disease (especially cancer) may be detrimental to some individuals, and that others may simply not want to know what the future holds. Furthermore, there are concerns that insurance companies may refuse to cover individuals with high-risk genetic profiles.

In summary, in the future, pharmacogenomics is expected to play a role in each step of the cancer management process including risk prediction and treatment.

< Prev   CONTENTS   Source   Next >