By Amit B. Agarwal, MD, PhD, and Christine M. Lovly, MD, PhD

	Amit B. Agarwal, MD, PhD Institution: The University of Arizona Cancer Center Specialties: Hematology, medical oncology Member since: 2011 ASCO activity: Professional Development Committee member and Career Development Committee Liaison of the Cancer Research Committee	Christine M. Lovly, MD, PhD Institution: Vanderbilt-Ingram Cancer Center Specialties: Thoracic malignancies, medical oncology Member since: 2008 ASCO activity: Professional Development Committee and Cancer Education Committee member

Genomic medicine can be transformative, but it also presents a set of unique challenges, both expected and unexpected.

The premise of genomic medicine

The term "genomic medicine" is used interchangeably with "personalized medicine" or "precision medicine," in both the medical literature and mainstream press. The description of the "central dogma" of molecular biology wherein information is passed from DNA to RNA to protein prompted researchers to find ways of studying this in the context of human diseases. The importance of the role of this highly complex and regulated process in cancers soon became clear. With the advances in available techniques to study these processes and the availability of drugs that can overcome imbalances in molecular pathways that result from these aberrations, genomic medicine has come to the forefront of cancer medicine. The premise of genomic medicine is that knowledge about the genomic changes that are a central characteristic of all cancers is driving not only our understanding of the complex molecular pathways but also improved therapeutic and diagnostic strategies.

Understand the available technology

The resolution of the tests available at our disposal has gone from looking at the entire chromosome using karyotyping to the ability to perform whole-genome sequencing in individual cells. In order to master genomic medicine, oncologists need to have a good understanding of the tests involved and the strengths and limitations of these techniques. For genes, where common mutations are present in certain "hotspots," a gene specific mutational analysis is considered optimal. Allele-specific polymerase chain reaction (PCR) is a commonly employed technique to genotype oncogenes such as BRAF, KRAS, JAK2, KIT, etc. Multiplex PCR techniques allow for simultaneous genotyping of a panel of different oncogenes. This system offers the advantage of a low-cost, easily accessible system that can be used on archival tumor specimens. The biggest limitation of this method is that an a priori knowledge of the mutation is required and uncommon mutations within the gene may not be tested.

Next-generation sequencing (NGS) techniques, on the other hand, allow for detection of the entire sequence of DNA. Massively parallel sequencing has revolutionized our ability to study molecular perturbations in tumor samples. The entire genome or specific subsets (exome, transcriptome) can be sequenced to obtain information about individual base pair variations for tumor samples. Significant costs and necessary expertise currently restrict NGS to research settings, but the decreasing cost is making it feasible for this technique to become part of the routine work-up of patients. In the coming years, we will have the ability to comprehensively study individual tumors for epigenetic, genetic, transcriptional, and translational perturbations to guide clinical decision-making. This information is beginning to inform various clinical decisions such as predicting sensitivity or resistance to certain drugs, predicting prognosis or risk of disease, and measuring responses to treatment and presence of minimal residual disease.

Drivers and passengers

The ability to sequence the entire genome from individual tumors has provided a great opportunity to identify all the molecular perturbations that are seen in the tumor cells, but a unique challenge has emerged. Not all molecular changes are central to the process of cancer initiation, maintenance, or progression. Several "passenger" mutations likely accrue stochastically in cancer cells but do not confer any growth advantage to the tumor and hence are not ideal targets for therapeutic development. "Driver" mutations, on the other hand, contribute to important steps in oncogenesis and are ideal targets. A common strategy to distinguish driver mutations from passenger mutations has been to study the frequency of aberration of a given gene in tumor types and compare that to the background mutation rate expected to occur randomly. With NGS, this task becomes quite daunting since passenger mutations far outnumber driver mutations and large-scale efforts such as The Cancer Genome Atlas (TCGA) project and International Cancer Genomics Consortium will be needed to help distinguish between drivers and mutations. Research efforts will have to then focus on defining biologic functions of novel putative driver mutations and ultimately developing therapeutic strategies to target molecular pathways perturbed by these drivers. In the meantime, critical thinking will be needed before designing clinical trials and in everyday practice to ensure that we are not chasing red herrings.

Application of genomic medicine

Recently, successful application of cancer genomics to develop therapies has led to U.S. Food and Drug Administration approval of several drugs, but some of the early success was seen at the turn of the century. The monoclonal antibody trastuzumab prolonged survival in breast cancer where the oncogene HER2 was noted to be amplified. This was followed by the inhibition of BCR-ABL in chronic myeloid leukemia using a small molecule tyrosine kinase inhibitor (imatinib), leading to dramatic outcomes. The results of vemurafenib in melanoma with BRAF V600E mutation and crizotinib in non-small cell lung cancer with ALK translocation highlight the potential and challenges of genomic medicine. Vemurafenib was approved for the treatment of late-stage melanoma in patients that harbor the BRAF V600E mutation in August 2011, only nine years after the mutation was first described in patients with melanoma. The drug prolonged survival in patients with the mutation, but despite dramatic initial response, patients have disease progression on the drug. A similar story has emerged for crizotinib, a small molecule inhibitor in patients with lung cancer with ALK translocation, except for an even shorter time from discovery of the perturbation to drug approval (four years). The resistance mechanisms to these first-generation drugs are the focus of intense research, and the next set of trials will look to overcome resistance by using second-generation inhibitors or drug combinations that target the involved pathway.

Ethical issues

With NGS, in addition to information about somatic mutations, a vast amount of information about germline variants is also generated. This information has implications for both individual patients and their family members—the handling of which raises ethical and legal issues. Particularly complicated is the issue of dealing with incidental findings. For example, what should an oncologist treating a patient for breast cancer do if whole-genome sequencing of her sample reveals that she has a variant that increases the risk of Alzheimer's dementia twofold as compared with the general population? Should this information be passed to her family if she dies? Since new knowledge becomes available every day, it is difficult to get an informed consent wherein the patient truly understands all possible implications of such tests. Another issue pertains to privacy. A recent report showed that individuals could be identified using a publicly available genomic database combined with other accessible information, thus compromising a person's identity. Going forward, all these issues will have to be considered.

The challenges ahead may be vast, but as fellows in oncology training, we have the opportunity to spearhead this changing paradigm and usher in the era of genomic medicine.