Genomic Messages: How the Evolving Science of Genetics Affects Our Health, Families, and Future - George Annas, Sherman Elias (2015)

Chapter 8. Cancer Genomics

Everyone who is born holds dual citizenship, in the
kingdom of the well and in the kingdom of the sick.

—Susan Sontag, Illness and Its Metaphors (1977)

Much of the action in contemporary genomics relates to cancer. There is a good reason for this. Cancer has been the most feared disease in the United States for more than a century. Efforts sponsored by the federal government to find cures for cancer date from the establishment of the National Cancer Institute (NCI) in 1937. Cancer research intensified after President Richard Nixon’s declaration of a “war on cancer” in 1971, and again in 2009 when President Obama announced he would double the amount of money the NCI could spend on cancer research, and arguably yet again in 2015 when President Obama proposed a new genomics research project to, among other things, cure cancer. In this chapter we report on both how we are doing in the genomic war against cancer and why curing cancer is so hard. An accurate description of the current state of genomic science and cancer treatment was made in Nature: “Personalized, ‘precision’ medicine for cancer is in a difficult time of transition.” It is a time of great promise and great challenge.

We begin with a homage to Susan Sontag, who after her first experience with cancer wrote a powerful meditation on the language we use to describe it, pointing out that the most common metaphors are military ones. As she observed, cancer cells don’t just multiply; they are “invasive,” they “colonize” the body, setting up outposts (“micrometastases”). The body’s “defenses” become overwhelmed, “scans” of the territory (the body) are taken, and treatment aims to “kill” the cancer. There is, Sontag writes, “everything but the body count.” She argued (mostly unsuccessfully) that we should abandon this metaphor because it is dehumanizing, leading us to treat the patient’s body like a battlefield, to set no limits in our attempts to destroy the enemy, and to place no cost limits on our efforts, no matter how futile.

Cancer is the second most common cause of death in the United States, exceeded only by heart disease. The average lifetime risk of a man developing cancer is slightly less than one in two; for women the risk is slightly higher than one in three. However, for any given individual, the risks vary depending on such factors as environmental exposures (for instance, smoking) and genetic susceptibility. Nearly one in four Americans will die of cancer. In 2014 there were more than 1.5 million new cancer cases and more than 500,000 cancer deaths.

Environmental factors (as opposed to hereditary factors) account for an estimated 75–80 percent of cancer cases and deaths in the United States. These include tobacco use, poor nutrition, physical inactivity, obesity, certain infectious agents, certain medical treatments, excessive sun exposure, and exposure to carcinogens (cancer-causing agents) that exist as pollutants in our air, food, water, and soil. Of these, tobacco smoking accounts for 30 percent of all cancers, and the combination of poor nutrition, obesity, and inactivity account for an additional 35 percent of the cancer burden. Environmental and behavioral factors are potentially modifiable. It has also been suggested that the majority of variation in cancers involving different sites, including esophagus, intestines, and stomach, are due to “bad luck,” specifically “random mutations arising during DNA replication in normal, noncancerous stem cells.”

Cancer Is a Genetic Disease

Although only 5 to 10 percent of cancers have a heritable component, cancer is a genetic disease, the most common of all genetic diseases. Cancer arises through mutations (changes in the sequence of DNA) leading to uncontrolled rapid growth of cells. We can divide these mutations into two types: somatic mutations (found in the cancer itself but not in the rest of the body’s cells) and germline mutations (inherited and present in every cell throughout the body). Most cancers are the result of somatic mutations, which are induced by environmental exposures. Nearly all cancers originate from a single cell through sequential accumulation of multiple mutations that ultimately overcome the control mechanisms regulating cell growth. The descendants of such a cell continue to divide, called clonal expansion, and undergo further genetic changes into invasive, metastatic cancer.

The purpose of cancer screening is to discover a malignancy at an early stage, before the patient has any signs or symptoms, so that treatment can be initiated, optimizing the chance of a cure. Cancer screening falls into five general categories: (1) history and physical examination (for example, unexplained weight loss or swollen lymph nodes); (2) laboratory tests (including blood, urine); (3) imaging studies (such as X-rays, ultrasound, or magnetic resonance imaging); (4) invasive procedures (for example, colonoscopy); and (5) genetic tests (mutations in specific genes).

All screening tests have disadvantages or carry risks. Invasive procedures, such as a colonoscopy, can rarely cause damage, such as perforation of the bowel. A test can show a false-negative result, in which the test indicates there is no cancer when there really is. This can lead to delay in treatment and adversely affect prognosis. A test can also show a false-positive result, in which the test is abnormal but cancer is not present. Beyond causing anxiety, this often leads to more invasive testing, such as biopsies, which result in possible discomfort, risk (such as bleeding and infection), and costs.

Some screening tests are offered to individuals who have a known increased risk for certain cancers (known as a risk factor). These include people who have had a prior cancer, those who have a strong family history of cancer, or those who have been identified as carrying gene mutations that predispose to cancer (usually discovered during the evaluation of a family history of cancer). The American Cancer Society provides guidelines for the early detection of cancer on their website.

A goal of genomic testing is to usher in an era of personalized cancer medicine. One idealized view of the future has been described by the American Society for Clinical Oncologists. In their hypothetical case, a woman (we’ll call her Joan) has a “routine” blood test during her annual physical, and within a few minutes the results come back as showing cancerous cells in her bloodstream. Joan is told that these cells “are an indication of an early-stage cancer that is developing somewhere in your body.” She is reassured that since the cancer has been detected at an early stage, “there is a good chance that it can be managed or cured.”

Joan is referred to an oncologist, who recommends additional genomic tests to determine the molecular “fingerprint” of the cancerous cells and the gene and protein abnormalities that may be driving her cancer. Within a few hours, the laboratory results are returned, indicating that Joan’s cancer is developing in her kidneys. The oncologist says it’s not the tumor’s location that’s important but her genomic profile and the unique combination of molecular features of her cancer. Based on Joan’s medical history and genomic predispositions in her electronic health record (EHR), as well as treatment information about other patients like her, the oncologist determines that Joan would probably have an adverse reaction to one of the standard therapies. The EHR also identifies a clinical trial in which she could enroll.

After obtaining a second opinion, Joan enrolls in the trial, which includes two new drugs, “which are attached to a microscopic ‘nanoparticle shuttle’ that will deliver them directly to individual cancer cells, sparing healthy cells and minimizing side effects.” Joan also receives a saliva reader that attaches to her smart phone and mobile applications that record her symptoms during the trial and automatically transmit this information to her EHR. Her smart phone also notifies her when it’s time to take her medication and asks her questions about how she is feeling. It also alerts her to expect to feel fatigued and provides suggestions for managing side effects. Joan feels reassured that her health care team knows a great deal about her cancer and that she is making informed decisions about managing her cancer, “while continuing to work and live an active life.” How close to reality is this futuristic vision of cancer diagnosis and treatment? To begin to answer this question, we must start with our current understanding of the causes of cancer.

The Hallmarks of Cancer

Cancerous tumors are complex tissues made up of malignant cells as well as normal cells that interact with and support the tumor. Therefore, to understand the biology of tumors, it is not enough to just study the malignant cells; rather, the entire “microenvironment” of the tumor must be examined. As normal cells evolve progressively to cancer, they acquire a succession of distinctive and complementary capabilities that enable the tumor to grow and spread. Douglas Hanahan, director of the Swiss Institute for Experimental Cancer Research, and Robert A. Weinberg of the Whitehead Institute for Biomedical Research at MIT have described these capabilities as eight “hallmarks of cancer” in one of the most widely cited papers in the scientific literature. Of these eight hallmarks, six are established hallmarks and two are so-called emerging hallmarks. They can be summarized as follows:

1. Sustaining proliferative signaling

Perhaps the most fundamental trait of cancer cells is uncontrolled stimulation of cell proliferation—that is, cell growth and division. This happens because cancer cells grow even when they are not receiving signals (messages) to grow. In other words, cancer cells behave as if a growth stimulus (called a growth “factor”) were present even when it is not. They are no longer dependent on growth factors; they become “masters of their own destinies.”

2. Evading growth suppressors

Most normal cells have a series of “brakes” that prevent them from continuing to grow and divide. Cancer cells lose their ability to respond to these “antigrowth signals,” which come from tumor suppressor genes.

3. Resisting cell death

Normal cells have a built-in mechanism for programmed “suicide,” termed apoptosis. For example, if a skin cell accumulates excessive DNA damage from too much exposure to UV light, it has mechanisms to self-destruct. Cancer cells have ways to disrupt the cell death mechanisms that carry out the suicide mission. Instead, cancer cells keep on growing and dividing.

4. Enabling replicative immortality

At the tips of chromosomes are stretches of DNA sequences called telomeres. Telomeres are made of a few thousand repeating sequences of TTAGGG. In normal cells, the telomeres limit the number of times a cell can divide. Each time a cell divides, the telomeres get shorter, in the range of ten to thirty-five TTAGGG repeats. Telomeres have therefore been compared to a fuse on a bomb. When the telomeres get too short (usually after fifty to seventy cell divisions), the cell can no longer divide, and it becomes inactive (“senescent”) or dies. By contrast, cancer cells develop ways to lengthen their telomeres, thereby allowing them to divide an infinite number of times. In this way cancer cells achieve a form of immortality.

5. Inducing angiogenesis

Like normal organs, cancerous tumors require blood vessels to provide oxygen and nutrients and dispose of carbon dioxide and wastes. Most tumors develop signaling mechanisms to switch on the growth of new blood vessels, called angiogenesis, to keep up with their expanding needs. These new blood vessels are usually abnormal: convoluted with excessive branching, enlarged, and having erratic blood flow and leakiness.

6. Activation of invasion and metastasis

As cancerous tumors develop into higher grades of malignancy, they invade local tissues and spread to distant locations, called metastasis. These cancer cells change in shape as well as their ability to attach to other cells and to the extracellular matrix (ECM), the meshwork that surrounds cells and organizes them into structures. These capabilities of cancer cells involve a multiple-step process of mutations in genes that encode molecules for cell-to-cell and cell-to-ECM invasion, the invasion-metastasis cascade.

7. Emerging hallmark: Reprogramming energy and metabolism

Cancers need to make adjustments to rapidly increase or redirect their energy metabolism (for example, usage of glucose) to fuel cell growth and division. This reprogramming of energy and metabolism is largely orchestrated by proteins that are involved in one way or another with the six aforementioned hallmarks.

8. Emerging hallmark: Evading immune destruction

There is an increasing body of evidence from animal and clinical studies suggesting that the immune system acts as a significant barrier to tumor formation and progression. It is thought that cells in at least certain types of cancer may evade destruction by disabling components of the immune system that have been dispatched to kill them.

Scientists are currently developing targeted therapies aimed at one or more of these hallmarks (figure 8.1). The sobering reality is that tumors seem to be able find a way around these targeted therapies, and cancers therefore often don’t stay in remission forever. The hope is that if multiple hallmarks are targeted simultaneously, a tumor won’t be able to get around all these barriers at the same time.


8.1  The Hallmarks of Cancer, D. Hanahan and R.A. Weinberg, “Hallmarks of cancer: the next generation,” Cell 144 (2011): 646-74.

Angelina Jolie Pitt is probably the best-known cancer patient in the United States, though, as we discussed in chapter 1, she has never had cancer. Americans like to read about celebrities, but they don’t like to read about dying celebrities. Nonetheless, the well-known acerbic and hard-living writer, Christopher Hitchens, famous for writing a nasty book trashing Mother Teresa, The Missionary Position, was determined to write about his own experience in cancer treatment. His story has special meaning for us because his progress was closely followed by NIH director Francis Collins (when you are a celebrity, you can get the attention of the head of the National Institutes of Health), who tried to come up with as many genomic treatment possibilities for Hitchens as he could. Hitchens wrote about his experiences in Vanity Fair, and the articles were later collected in his book Mortality. Hitchens, a long-time smoker and heavy drinker, had esophageal cancer. He describes his experience: “I have more than once in my time woken up feeling like death,” but this time was different; he could barely breathe, and it took all his strength to call for emergency personnel. Thinking back, he found them courteous and professional as they escorted him, “a very gentle and firm deportation, taking me from the country of the well across the stark frontier that marks off the land of malady.” In the emergency room, he was told his “next stop would have to be with an oncologist.” He describes his new citizenship in “the sick country” as having some advantages: “The new land is quite welcoming in its way. Everybody smiles encouragingly and there appears to be no racism. . . . [A]s against that, the humor is a touch feeble and repetitive, there seems to be almost no talk of sex, and the cuisine is the worst of any destination I have ever visited.”

He is informed that the cancer had spread to his lymph nodes and that they are “palpable” from the outside; he would get biopsy results in a week. Eventually Hitchens gets settled into a regime of chemotherapy and becomes philosophical, noting that the “absorbing fact about being mortally sick” is that you spend a lot of time preparing to die “while being simultaneously interested in the business of survival.” In one of our favorite phrases, he writes, “This is a distinctly bizarre way of ‘living’—lawyers in the morning and doctors in the afternoon—and means that one has to exist even more than usual in a double frame of mind.”

A well-known atheist, Hitchens writes about all the believers supporting him. The “best of the faithful,” he writes, is Francis Collins, whom he knows both for his genome work and “from various public and private debates over religion.” Collins “has been kind enough to visit me and to discuss all sorts of novel treatments, only recently even imaginable, that might apply to my case,” although later he writes, “In Tumortown you sometimes feel that you may expire from sheer advice.” Hitchens is encouraged to explore a new “immunotherapy protocol” being done at NCI, which involves removing T cells from the blood, subjecting them to genetic engineering, and then reinjecting them to attack the cancer cells. There is a catch: his tumor cells have to “match,” they have to express a protein called NY-ESO-1, and his immune cells have to have the molecule HLA-A2. His tumor has the protein but not the HLA match. Other trials are said to be under way, but he writes, “I am in a bit of a hurry, and I can’t forget the feeling of flatness that I experienced when I received the news.” Later, when 60 Minutes runs a story of tissue engineering used to create a replacement esophagus for a man with his cancer, he immediately contacts Francis Collins, “who gently but firmly told me that my cancer has spread too far beyond my esophagus to be treatable by such means.”

Realizing it’s a very long shot, Hitchens decides to have “my entire DNA ‘sequenced,’ along with the genome of my tumor.” Francis Collins picks up the story at this point, writing to Hitchens that if both sequences are done, “it could be clearly determined what mutations were present in the cancer that is causing it to grow. The potential for discovering mutations in the cancer cells that could lead to a new therapeutic idea is uncertain—this is at the very frontier of cancer research right now” (emphasis added). Yes, all these new technologies might help, but they came too late for Hitchens, who deteriorated quickly and died on December 15, 2011. Two of his last written lines: “Body turns from reliable friend to more neutral to treacherous foe . . . Proust?” and “Banality of cancer. Entire pest-house of side-effects. Special of the day.”

Cancer-Genome Sequencing

The emergence of massively parallel sequencing (MPS), sometimes referred to as “next-generation sequencing,” has been a major technological advance used in cancer-genome sequencing, the one Christopher Hitchens was hoping to take advantage of in 2011. MPS refers to a group of so-called high-throughput DNA-sequencing technologies, meaning that they can rapidly sequence DNA on the hundreds of gigabase (1 billion bases) scale. This has enabled whole cancer genomes to be sequenced on large sample sizes of tumor types. It has also made it possible to use a single technology to perform many kinds of genome analysis—for example, discovering single-point mutations (that is, a change of one base pair in the DNA sequence), assessing copy number alterations (an abnormal number of copies of one or more sections of the DNA), and detecting DNA methylation (a method by which gene expression is regulated) and foreign DNA (such as from viruses).

Bert Vogelstein and colleagues from Johns Hopkins University have noted that cheaper and more accurate genome-sequencing technologies are rapidly changing our fundamental understanding of cancer. The major points are that many genes are mutated in a typical cancer (in common solid tumors, such as those originating from the colon, breast, brain, or pancreas, an average of thirty-three to sixty-six genes show mutations not found in the individual’s noncancerous cells), and like snowflakes, no two cancer genomes are alike. Tumors evolve from benign to cancerous by acquiring a series of mutations over time; some mutations, called “driver mutations,” affect cellular signaling and regulatory pathways, each containing multiple genes.

There are over two hundred forms of cancer, and each type has its own genetic landscape. In 2006 the National Cancer Institute and the National Human Genome Research Institute launched the Cancer Genome Atlas (TCGA) to systematically map these genetic changes. Investigators from around the world sent normal and malignant tissue samples from patients to the TCGA researchers, up to five hundred samples for each tumor type. The cancer types were selected for their poor prognosis and overall public health impact. DNA, protein, and RNA were extracted from the samples, and each type of cancer was extensively characterized. The data has been made freely available to cancer researchers worldwide. The TCGA was declared completed at the end of 2014, but the NCI has stated that it will continue intensive sequencing in three cancers: ovarian, colorectal, and lung. Plans for the new studies call for including clinical data about each patient as well.

A frequent characteristic of cancer cells is that their nuclei are enlarged, distorted, and darkly stained with clumping; sometimes their appearance is described as “bizarre.” The reason for this appearance is that the nucleus of a cancer cell often contains major chromosomal abnormalities, namely too many chromosomes. Another type of chromosomal abnormality frequently seen in cancer cells is a “translocation,” where a piece of chromosome breaks off and sticks to another chromosome. The most common type of translocation in cancer cells is “reciprocal translocation,” when pieces of two chromosomes exchange places. Most solid tumors have dozens of translocations. However, the majority of translocations appear to be passengers rather than drivers, because their breakpoints occur in “gene deserts,” parts of chromosomes that have no known genes.

The textbook model of tumor development is a “Darwinian” competition whereby driver mutations enhance a cell’s “evolutionary fitness,” promoting outgrowth of that clone and progression toward cancer. Genome-wide sequencing studies of several thousand tumors have shown only about two hundred common driver genes. Of these, about half are tumor suppressor genes—genes that normally prevent cell growth and proliferation but mutate and stop working. The other half are oncogenes, genes that normally promote cell growth and proliferation, and have the potential to become hyperactive. Although additional mutation driver genes will undoubtedly be discovered, they will likely be in uncommon tumor types that have not yet been studied in detail. The acquisition of driver mutations is gradual and occurs cumulatively over years to decades.

Cancer Treatment Is Evolving

As we discussed in chapter 3, personalized or precision medicine has been one of the most hyped “revolutions” in modern medicine, and nowhere has the hope for its success been greater than for cancer. For over half a century, the mainstay of cancer treatments beyond surgery has been the use of chemotherapy drugs and radiation treatments that attack all fast-growing cells, killing those that normally grow fast (like white blood cells or cells that line the intestines) as well as the cancerous ones. In essence, this is a one-size-fits-all approach to cancer treatment. By contrast, personalized cancer medicine uses specific genomic information about a person’s tumor to help establish a diagnosis, plan treatment, find out how well treatment is working, or make a prognosis. In broader terms, personalized genomic medicine takes into account both the tumor and the host—that is, genetic variations within the tumor (somatic mutations) and variations in normal tissues (germline mutations).

Genomic information about tumors helps in selecting an effective therapy and, equally as important, avoiding ineffective treatments. Inherited differences in how drugs are absorbed, metabolized, distributed in various tissues, and excreted can play a critical role in establishing proper chemotherapy dosing and staying clear of treatments with unacceptable risks of adverse drug effects (see chapter 5). Focused profiling of tumor DNA paired with anticancer drugs is increasingly being considered part of routine cancer care. For example, in 2014 the FDA approved a new drug for recurrent ovarian cancer called olaparib, together with a companion diagnostic test for specific mutations in the BRCA genes that would determine which patients were most likely to benefit from the new drug. There are now freely available online personalized cancer medicine resources, such as My Cancer Genome and the National Comprehensive Cancer Network, which serve as decision-making tools for physicians, patients, and researchers, providing up-to-date information on what mutations make cancers grow and related therapeutic implications, including available clinical trials.

Steve Jobs, cofounder of Apple Computers, provides an example of the role and limitations of cancer genomics. President Obama called Jobs “among the greatest American innovators—brave enough to think differently, bold enough to believe he could change the world, and talented enough to do it.” Jobs was one of the wealthiest people in America, with an estimated net worth of $7 billion. In 2003, Jobs had a CT scan and other tests, which determined he had a rare form of pancreatic cancer, an islet-cell neuroendocrine tumor. Such tumors arise from the specialized cells within the pancreas that produce insulin, which controls blood glucose. His physicians urged him to undergo surgery to remove the tumor, but instead Jobs chose a variety of alternative remedies, including a vegan diet, juices, herbs, and acupuncture. Nine months later, the tumor continued to grow. Only then did Jobs agree to surgery and chemotherapy.

According to an authorized biography, Steve Jobs by Walter Isaacson, Jobs personally led his doctors on an aggressive scientific approach to the treatment of his cancer. He was one of twenty people in the world at the time to have all the genes of his cancer and normal DNA sequenced. The sequencing cost $100,000 and was done through a collaborative effort among Stanford, Harvard, Johns Hopkins, and the Broad Institute of MIT. Based on the unique molecular signature of Job’s tumor, his doctors tailored his chemotherapy treatments. Jobs told Isaacson that he was either going to be the first “to outrun a cancer like this” using genomic sequencing technology or among the last “to die of it.” Jobs died on October 5, 2011, at age fifty-six, two months before Christopher Hitchens died. The cancers that caused the deaths of Hitchens and Jobs were relatively rare. We now turn to two more frequent cancers that have been especially prominent in genomics research: skin cancer (melanoma) and breast cancer.


Overall, the lifetime risk of developing melanoma skin cancer is about 2 percent (one in fifty) for whites and about 0.1 percent (one in a thousand) for blacks. This translates to about 77,000 new cases of melanoma diagnosed and about 9,500 people dying of melanoma each year. It is one of the most common cancers in young adults, especially women. The rate of melanoma has been rising for at least thirty years and now accounts for about 5 percent of all new cancers. Patients with early-stage melanoma can usually be treated successfully with surgical removal, although the disease can spread widely. The primary treatment for patients with metastatic melanoma has been standard chemotherapy. The side effects of these drugs are highly toxic, including severe nausea, anemia, and hair loss. In the past, the prognosis for patients with metastatic melanoma was very poor, with median survival rates well under one year.

Historically, melanoma has been classified based on clinical characteristics like the thickness of the primary tumor, the number of cells dividing in a biopsy specimen, and whether there are ulcerated lesions. In recent years, however, it has become clear that the genomic makeup of melanoma tumors has important therapeutic implications. It is now known that about half of melanomas carry mutations in a gene called BRAF. (BRAF is the awkward acronym for serine/threonine-protein kinase B-Raf.) Up to 90 percent of the mutations in the BRAF gene affect just a single DNA base change of a T to an A (called the BRAF V600E mutation). In turn, this change in the gene’s DNA sequence results in substitution of the amino acid valine for another amino acid, glutamic acid, in the BRAF protein. The BRAF protein helps transmit chemical signals from outside the cell to the cell’s nucleus. It plays an important role in one of the signaling pathways, known as the MAPK (short for mitogen-activated protein kinase) pathway. The MAPK pathway helps control cell proliferation, differentiation, movement (migration), and apoptosis (cell death). When the BRAF V600E mutation occurs, it disrupts normal functioning of the MAPK pathway; melanoma cells containing mutant BRAF disrupt MAPK signaling, leading to unbridled growth and survival.

A study published in 2010 involved thirty-two patients whose metastatic melanomas were unresponsive to standard therapy. DNA analysis of their tumors showed that all had the BRAF V600E mutation. These patients were given an orally administered drug, PLX4032 (later named vemurafenib), which specifically inhibited the BRAF protein made by the defective gene. Remarkably, this BRAF inhibitor induced complete or partial tumor regression in 81 percent of the patients and symptom improvement in the majority of patients. In some cases, lesions that had spread to liver, bone, and lungs seemed to disappear within weeks of the first dose. An accompanying editorial, entitled “Melanoma—an Unlikely Poster Child for Personalized Cancer Therapy,” exclaimed that “these results represent a major breakthrough and provide proof of principle that the treatment of metastatic melanoma can be individualized for a substantial percentage of patients.”

Confirming reports soon followed. Patients with the BRAF V600E mutation treated with vemurafenib or other BRAF inhibitors showed high and rapid response rates. Although side effects were common, they were generally not severe. Unexpectedly, however, about 25 percent of patients treated with BRAF inhibitors developed another type of skin cancer, squamous cell carcinomas. These lesions could usually be managed successfully by local surgical removal and did not require discontinuation of BRAF inhibitor therapy.

Then bad news started to come in. Responses were mostly partial, and about half of patients began having cancer progression within six to seven months of treatment, as the tumors became resistant to the BRAF inhibitors. Inhibiting BRAF activity alone would not eradicate or even hold back melanoma tumors for very long. A number of new strategies are being tried. For example, combining a BRAF inhibitor with an inhibitor of a second enzyme in the MAPK pathway, called MEK (mitogen-activated protein kinase), helps fight drug resistance. Combining inhibitor drugs within and across signaling pathways, which target the molecular makeup of an individual’s cancer, holds real promise in the treatment of melanoma as well as other cancers.

Another type of personalized cancer treatment, called “immunotherapy,” attempts to target specific cancer cells by identifying proteins on their surface encoded by their unique genes. Ways to unleash the power of the body’s own (“personal”) immune system on cancers have long been sought. The origins of modern immunotherapy for cancer can be traced back to the late nineteenth century, when William Coley, a young New York surgeon, began injecting live or inactivated bacteria into patients with cancer. Inoculating cancer patients with bacteria made sense, given the evolving understanding of the power of the body’s immune system to cause inflammation and destroy invading bacteria by stimulating antibacterial white cells that might kill bystander tumor cells. “Coley’s toxin” was used to successfully treat a wide variety of cancers, including melanoma, sarcomas, carcinomas, and lymphomas; complete, prolonged regression of advanced malignancy was documented in many cases. Use of Coley’s toxin was nonetheless opposed by the medical establishment because of mixed results. The modern science of immunology has shown that Coley’s principles were correct, and Coley is now known as the “father of immunotherapy.”

There have been intensive efforts to develop immunotherapeutic approaches to cancer. This is no easy task, because cancers have complex ways of escaping immune detection. For instance, tumors can decrease the expression of proteins on their surface, rendering them invisible to certain types of white blood cells (called cytotoxic T cells) that destroy target cells. Some tumors secrete proteins that actively suppress antitumor immune responses. Nonetheless, advances have been made in cancer immunotherapy. It has been observed that the “best approach to treatment might be to combine precision [genomic] therapy with immunotherapy.” The thought is that genomic-based interventions can provide a short-term fix, whereas immunology approaches, when they work, are effective for much longer periods of time.

One strategy that has been used in melanoma and other cancers involves a gene called PDCD1, which encodes for a protein called PD-1 (short for programmed cell death protein 1). PD-1 is a receptor on the surface of T cells that interacts with a molecule called PD-L1 (short for programmed cell death 1 ligand). Normally PD-L1 fits with PD-1 like a lock and key to maintain the balance of the immune system by shutting it down at appropriate times. This is called an “immune checkpoint.” Some melanomas (and other cancers) take advantage of this shutdown mechanism by continuously producing PD-L1, enabling the cancerous cells to escape being destroyed by T cells.

The drug pembrolizumab (formerly lambrolizumab) is a highly selective antibody against PD-1 that blocks the cancer cells from turning off the immune regulatory signaling of the PD-1 receptor expressed by T cells. This blocking antibody enables T cells to recognize proteins on cancerous cells as “foreign.” The antibody doesn’t fight the cancer directly, but it allows the body’s own immune system to target cancer cells for destruction. In a clinical trial of 135 patients with advanced metastatic melanoma who were treated with lambrolizumab, 38 percent improved. Among those who received the highest dose, 52 percent improved and 10 percent had a complete response, meaning their tumors could no longer be detected on scans. Side effects were mild and easily managed. A second clinical trial showed comparable results “with rapid and deep tumor regression in a substantial proportion of patients.” The results of these two clinical trials were heralded as “striking” and described as taking immunotherapy for melanoma to the next level.

Ralph Steinman, former director of the Laboratory of Cellular Physiology and Immunology at Rockefeller University, is an example of the potential of immunotherapy. In 1973, Steinman and a colleague, Zanvil Cohn, discovered a new class of cells, known as dendritic cells, that directs and regulates the body’s immune system by programming other cells to recognize and destroy intruders. In 2007, Steinman was diagnosed with advanced pancreatic cancer. With the help of dozens of collaborators around the world, he began running a series of research studies on himself, a kind of personalized immunotherapy. In all, he tried eight experimental treatments, each approved under a single-patient, compassionate-use protocol. Steinman used three vaccines, all based on his dendritic-cell research. The idea was to boost his own immune response to the cancer by inserting proteins from the surface of his tumor into the dendritic cells and injecting them back into his body. Survival for patients with Steinman’s type and stage of pancreatic cancer is usually measured in weeks or months; he lived for more than four years. Which, if any, of his treatments extended his life is unknown, but the work that he and his collaborators did advanced the field by demonstrating that conventional chemotherapy could be combined with dendritic-cell vaccines.

On October 3, 2011, it was announced that Steinman was to share in the Nobel Prize in Physiology or Medicine. Hours after the announcement, the committee learned that Steinman had died three days earlier. Deceased individuals have not been eligible for the prize since 1974. The committee consulted with their lawyers, who interpreted the rules as not applying if the committee was unaware that the winner was dead. Like Steinman’s work on immunotherapy, change in American medicine is not revolutionary but evolutionary. Nowhere is this truer than in the area of cancer. We all hope for major cures, so-called magic bullets that will take away or, better yet, prevent cancers in ourselves and our loved ones. Such “overnight” breakthroughs rarely happen, but personalized medicine is slowly taking hold.

Breast Cancer

In the United States, about 230,000 persons will be diagnosed with invasive breast cancer, and about 40,000 deaths will occur from the disease in 2015, assuming current trends continue. One in eight women born in the United States will develop breast cancer sometime before they die. Another way of saying this is that one in eight women in the United States who reach the age of 80 can expect to develop breast cancer—although this figure is misleading and includes cancers that will never be detected. In each decade of life, the risk of getting breast cancer is actually much lower than one in eight, although risk increases with age. For example, the ten-year risk for breast cancer is one in sixty-nine (or less than 2 percent) for a woman at age forty; it is one in forty-two (a little more than 2 percent) at age fifty, and one in twenty-nine (about 3 percent) at age sixty. These probabilities are averages for the whole population. An individual woman’s breast cancer risk may be higher or lower depending on a number of factors. The benefits and harms of a fifty-year-old woman being screened with an annual mammogram each year for ten years (until she’s sixty) are illustrated in figure 8.2. As seen, the woman will have more than a 60 percent chance of at least one false positive, 940 (or 9.4 percent) will get an unnecessary biopsy, fifty-seven will be overdiagnosed, sixty-two will die regardless of screening, and ten deaths will be averted. Only one woman in three will have a normal mammogram each year for the ten-year period. The bottom line is not that mammography should be discontinued for this age group, only that it helps postpone very few deaths (10), and six times as many breast cancer deaths (62) occur despite yearly mammography. Of course in the nondeath category, 940 will get unnecessary biopsies, 57 will get overdiagnosed and undergo unnecessary treatment, and 173 will survive breast cancer regardless of screening. Few clinicians, and fewer women, will disagree that we need a more effective approach to breast cancer prevention. Family history can provide a more effective approach for women at risk for heritable breast and ovarian cancer.


8.2  Visual aid to understanding breast cancer screening. J. Jin, “Breast Cancer Screening: Benefits and Harms,” JAMA 312 (2014): 2585.

Diagnostic evaluation is usually undertaken because the patient or the health care professional feels a suspicious breast lump, or there is a suspect finding on a screening mammogram. Additional imaging studies using ultrasonography or magnetic resonance imaging (MRI) may be used. The next step is to obtain cells or tissue for diagnostic studies from the suspicious lesion.

Once breast cancer is found, a number of tests are performed to classify the tumor, determine the prognosis, and select the optimal treatment. Breast cancer is commonly treated by various combinations of surgery, radiation, chemotherapy, and hormone therapy. Important factors that need to be assessed to provide a personalized approach to cancer treatment include how quickly the cancer is likely to grow, how likely it is to spread throughout the body, how well certain treatments might work, and how likely the cancer is to recur.

After gender and age, a positive family history is the strongest known predictive risk factor for breast cancer. About 10 percent of breast cancers are linked to germline mutations, due to a hereditary mutation that is passed down from parent to offspring. Chief among them is the hereditary breast and ovarian cancer syndrome. Characteristics of this syndrome are early onset of breast or ovarian cancer (usually defined as younger than age forty) in multiple generations, individuals with cancers developing in both breasts or ovaries, Ashkenazi Jewish descent, and sometimes the occurrence of other cancers in family members (for example, prostate, pancreas, uterus, colon). Some of these families (some writers call them “cancer families,” but it is wrong to reduce entire families to one characteristic they share) can be explained by single cancer susceptibility genes, with germline mutations in the BRCA1 and BRCA2 genes accounting for the vast majority of families with hereditary breast and ovarian cancer syndrome. The names BRCA1 and BRCA2 stand for breast cancer susceptibility gene 1 and breast cancer susceptibility gene 2, respectively.

The BRCA1 gene is located on chromosome 17 and the BRCA2 gene on chromosome 13. Both BRCA1 and BRCA2 are tumor suppressor genes encoding for proteins that function in the DNA repair process (discussed above). More than 2,500 distinct mutations have been described in BRCA1 and BRCA2. In the general population, about 1 in 300–800 individuals carry a BRCA1 or BRCA2 mutation. About 3–5 percent of breast cancers and 10 percent of ovarian cancers are due to germline mutations in BRCA1 and BRCA2. In the United States, about 1 in 40 individuals of Ashkenazi Jewish descent carry two specific BRCA1 mutations (designated 185delAG and 538insC) and a BRCA2 mutation (designated 617delT).

The lifetime estimated risk of a woman developing breast cancer ranges from 65 to 78 percent for BRCA1 or BRCA2 mutation carriers; the lifetime risk of a woman developing ovarian cancer ranges from 39 to 46 percent for BRCA1 and 12 to 74 percent for BRCA2 mutation carriers. Men with BRCA2, and to a lesser extent BRCA1, are at increased risk of breast cancer with lifetime risks in the range of 5–10 percent and 1–2 percent, respectively. Men carrying BRCA2 mutations, and to a lesser extent BRCA1 mutations, have about a three- to sevenfold increased risk of prostate cancer.

Of course, not every woman who carries a BRCA1 or BRCA2 mutation will develop breast or ovarian cancer. Even at a 78 percent risk, 22 percent, or almost one in four, will never develop the disease. Moreover, not every woman in families with hereditary breast and ovarian cancer syndrome necessarily carries a BRCA1 or BRCA2 mutation, and not every cancer in such families is necessarily linked to a harmful mutation of one of these genes. Evaluating a patient’s risk for hereditary breast and ovarian cancer syndrome should be part of routine clinical practice. A number of clinical practice guidelines provide specific criteria clinicians may use in determining who should be offered genetic testing for BRCA1 and BRCA2 mutations and whether referral for genetic counseling is appropriate. Genetic counseling is generally recommended before and after genetic testing. The National Cancer Institute’s recommendations for BRCA1 and BRCA2 testing are on its website.

A positive genetic test result for a BRCA1 or BRCA2 mutation has important health and social implications for carriers, other family members, and future generations. When evaluating a family, it is best, if possible, to begin with a person who has developed breast or ovarian cancer. Once a specific mutation has been identified in the affected individual, others in the family who choose to be tested can be studied for the same mutation. It is important to remember that both men and women who inherit a BRCA1 or BRCA2 mutation, whether or not they develop cancer themselves, can pass the mutation to their sons or daughters. There is a one in two chance of a mutation carrier passing the abnormal gene to a child. Conversely, there is a one in two chance that a mutation carrier will not pass the abnormal gene to a child.

If a person tests negative for a known mutation in his or her family, it is highly unlikely that they have inherited a susceptibility to the breast and ovarian cancer syndrome associated with BRCA1 or BRCA2. However, this does not mean that they will not develop cancer; it means that the person’s risk is probably the same as for anyone in the general population: one in eight. However, where there is a family history of breast and ovarian cancer but no mutation in BRCA1 or BRCA2 is identified, a negative test is considered “not informative.” In other words, a harmful BRCA1 or BRCA2 genetic abnormality could exist, but it was not detected by the test. In addition, it is possible for people to have a mutation in a gene other than BRCA1 or BRCA2 that increases their risk of developing cancer but is not detectable by the tests used.

The recommendation to only test for cancer genes in healthy people if a close family member has been found to carry a particular gene was called into question in late 2014 with the publication of a study in Israel that found women of Ashkenazi backgrounds have an unusually high incidence of BRCA1 and BRCA2 genes. While such screening is not currently recommended, many other “cancer genes” can now be screened for, with various panels of genes being suggested by different biotech companies. The quandaries these panels present to physicians and patients are well described by Science writer Jennifer Couzin-Frankel, who wrote about her own experiences (figure 8.3). Like many others, she read the New York Times article about the Israeli study in September 2014. Because both her parents were of Ashkenazi descent, she decided to consult a genetic counselor about BRCA1 and BRCA2 screening (the risks of carrying a BRCA mutation is one in eight hundred in the general population, and one in forty in the Ashkenazi Jewish population). After reviewing her family history, which included some possible cancers, the genetic counselor presented a list of twenty-one genes associated with breast and ovarian cancer, the genes on the “Breast/Ovarian Cancer Panel” marketed by a company called GeneDx in Maryland. As she described the table, eleven genes were shaded in pink and labeled “high-risk”; three in purple were “moderate-risk”; and seven in turquoise were described simply as “newer genes.” Other companies recommended other genes. The CHEK2 gene, colored purple, for example, could, she was told, double her risk of breast cancer. The impact of most of these genes on health remains uncertain—but as experts she consulted with (and had access to because of her profession) told her, the ability to test is moving much faster than our ability to interpret test results. As Kenneth Offit of Sloan Kettering Cancer Center told her, “This is the paradox we have fallen into.” We are discovering more cancer genes at the precise time the cost of genomic sequencing is plummeting, with the result that the number of testing panels is proliferating.”


8.3  Science writer Jennifer Couzin-Frankel and her children. April Saul, in J. Couzin-Frankel, “Unknown Significance” Science 346 (2014): 1167.

In the article, which we recommend you read in its entirety, she describes her decision-making process, as well as her own tolerance for uncertainty, including whether she’d want to be informed of variants of uncertain or unknown significance. Ultimately she decides to have the panel of tests and to be informed of all the results, regardless of their known significance. She learns there are no variants of unknown significance in BRCA, but there is one in CHEK2. She is told that the latter variant had been found in two men with prostate cancer. Her response: “I expected distress, a ringing in my ears, fear coiling in the pit of my stomach. Instead, I’m almost laughing. I think, ‘That’s it? That’s what is being shared with patients these days?’ Two men with prostate cancer, cells in a petri dish, a loss of function that may or may not translate into pathogenicity: This does not merit my mental energy.”

In the end she shares her results with her cousin, the only family member to whom it might matter. Her cousin urges her to get tested with a new panel—one for forty-eight genes, a test her cousin’s mother just had because of her ovarian cancer (all negative). “In the end, I explain in my message to her, it wasn’t something I wanted. ‘I know the panels are often discouraged,’ her cousin writes back. It’s a view she doesn’t share. Even without a clear-cut action plan, she wants to know whatever message her DNA carries for her future. The only reason she’s eschewed testing for herself is because insurance is unlikely to pay for it.”

Finding variations of uncertain significance is not unusual, even when testing is restricted to BRCA1 and BRCA2. Overall, about 10–15 percent of individuals undergoing screening for BRCA1 and BRCA2 mutations will not find a clearly harmful mutation but will have a variant of unknown (or uncertain) significance (VUS). However, the proportion of individuals who receive a VUS result varies widely by population, with rates of up to 22 percent among Hispanics and 26 percent among African Americans. A VUS may cause substantial challenges in counseling in terms of cancer risk estimates, and clinical decision making must be highly individualized and take into account factors such as the patient’s personal and family cancer history. Moreover, inherent in the implications of VUS is that they are moving targets. As additional data becomes available, a VUS may be reclassified as a benign variant or as deleterious.

Individuals who are identified as carriers of a harmful BRCA1 or BRCA2 mutation have a number of options to reduce their risks of developing cancer. These include:

• Surveillance. Surveillance refers to cancer screening with the goal of diagnosing the cancer early, when it is most treatable. Breast cancer screening may include mammography, clinical breast exams, and other breast cancer screening methods, such as MRI. While no specific type of surveillance has been shown to improve patient outcomes for those at high risk for ovarian cancer, some patients have opted for close monitoring using blood tests (for instance, CA125), transvaginal ultrasounds, and clinical evaluations. The effectiveness of such screening methods is uncertain.

• Prophylactic surgery. Bilateral mastectomy reduces the risk of breast cancer by 95 percent or more. Bilateral salpingo-oophorectomy (removal of both ovaries and fallopian tubes) has also been shown to reduce the risk of breast cancer by 40–70 percent. Bilateral salpingo-oophorectomy reduces the risk of ovarian cancer, fallopian tube cancer, and peritoneal cancer by 90 percent or more. Unfortunately, because not all “at-risk” tissue can be removed by these procedures, some women have developed breast cancer, fallopian tube cancer, ovarian cancer, or peritoneal cancer even after prophylactic surgery.

• Chemoprevention. Some breast cancers can be prevented by limiting cellular access to estrogen, which stimulates cell growth. Estrogen receptor blockers, such as Tamoxifen and Raloxifene, have been shown to reduce the risk of developing breast cancer in BRCA1 and BRCA2 mutation carriers.

The story of Ms. E (a real patient whom we’ll call Emily) exemplifies how personalized cancer medicine is currently being practiced in the United States. Emily is a forty-one-year-old woman who works in the banking business. She is married and has had one child. At age thirty-seven, she began experiencing pain in her left armpit and noticed a lump that enlarged slightly over the next six weeks. She saw her gynecologist, who ordered a mammogram that showed only dense breast tissue, but MRI and ultrasound examinations revealed a mass in her breast as well as suspicious lymph nodes. A biopsy was taken of the mass, which showed an invasive ductal carcinoma of the breast. Emily was treated with wide surgical excision of the tumor and chemotherapy with three drugs (doxorubicin and cyclophosphamide followed by paclitaxel) and radiation therapy.

Because of her young age, Emily was referred to genetic counseling and underwent testing that showed she had a BRCA1 mutation, more precisely a 538insC BRCA1 mutation, which is one of the two common BRCA1 mutations found among individuals of Ashkenazi Jewish descent. She was unaware of any Ashkenazi Jewish ancestry; however, her father was an adopted only child. In addition, the pathological studies performed on the biopsy tissue showed that her tumor was triple negative (ER-/PR-/HER2-), which is the case in 80 percent of BRCA1-related breast cancers. This was important information, because although recent data has suggested that an unconventional chemotherapeutic drug for breast cancer, cisplatin, might be useful for the treatment of newly diagnosed breast cancers in BRCA1 mutation carriers, the standard therapy that Emily received was believed to be more appropriate than cisplatin therapy outside of a clinical trial.

Additionally, it is currently recommended that mutation carriers like Emily undergo prophylactic removal of the ovaries and fallopian tubes when childbearing is completed, ideally by age thirty-five to forty, to substantially reduce the risks of ovarian cancer and breast cancer. However, at the time of her diagnosis, Emily and her husband told her oncologist that they were trying to have a child through in vitro fertilization. Emily’s eggs were harvested before she began chemotherapy. Subsequently, preimplantation genetic diagnosis was performed and an embryo free of the BRCA1 mutation was implanted into Emily’s uterus. (See chapter 4 for further details about preimplantation genetic diagnosis.) Four years after the initial diagnosis of her breast cancer, Emily gave birth to a healthy boy. Emily was quoted as saying, “The results of the genetic test have been extremely important for us. The idea of knowing is so much more important than not knowing. For us to have children at this stage in life and understand that we are not passing the gene outweighs any possible negative consequences.”

A Final Perspective

What can we conclude from our exploration of cancer and the genome? All cancers arise from changes in the DNA sequence of normal cells. We have entered an era in which genomic profiling has begun to significantly improve the treatment outcomes for some patients with cancers. Importantly, some of the new drugs developed from what has been learned about the molecular biology of any one cancer, for example melanoma, hold promise in targeting pathways and checkpoints in other malignancies, such as colorectal carcinoma and kidney cancer. Still, we have a long way to go. As Frances Visco, the president of the National Breast Cancer Coalition, put it in the spring of 2015, “Having the technology to do genomic analysis of tumors does not mean that we have ‘precision’ or ‘personalized’ medicine,” or that we can better help people with cancer. We agree with her that progress in cancer genomics must be measured not simply by developing “a new tool but rather stopping cancer deaths.”

You and your family will increasingly encounter genomics in everyday health care, and screening for cancer or for a predisposition to develop cancer will likely play a prominent role. In the context of “personalized medicine,” this will involve adjusting each person’s level of risk rather than the average population risk. For example, men whose genomic profile show a lower than average risk for dying of prostate cancer may be able to avoid screening with prostate-specific antigen (PSA) with its inherent high false-positive rate, which leads to unnecessary biopsies.

We can now foresee a time, realistically as soon as a decade from now, when we will look back on the ways we now diagnose, treat, and try to prevent cancer as a kind of dark ages. As Siddhartha Mukherjee put it in The Emperor of All Maladies, “Arguably [the] most complex, new direction for cancer medicine is to integrate our understanding of aberrant genes and pathways to explain the behavior of cancer as a whole, thereby renewing the cycle of knowledge, discovery, and therapeutic intervention.” With apologies to Susan Sontag, who might nonetheless agree with our conclusion if not our metaphor, the war on cancer has gone on far too long and has had far too many casualties. The peace treaty will be written in the language of our genomes and the genomes of the cancers that afflict us.

Further advances in cancer genomics will require sequencing hundreds of thousands, if not millions, of genomes so that relevant pathological genetic variants can be distinguished from irrelevant or harmless ones. This massive research project will require developing massive genomic data banks, and it is to the development of these data banks, and the privacy protections needed to encourage people to donate their genomes to such data banks, that we turn to in the next chapter.




Cancer is a genomic disease that usually
has an environmental trigger but can also
be the result of just plain bad luck.

The “hallmarks” of cancer suggest
specific components in cancer cells that
can be targeted for treatment.

Sequencing cancer genomes holds
promise for personalized medicine.

Immunotherapy is a fast-growing approach to
cancer treatment that can be applied alone or
in combination with genomic approaches.

Genomic screening for cancer can
increase uncertainty by identifying
variants of uncertain significance.

Breast cancer screening by mammography
illustrates some of the problems of
false positives and “overdiagnosis” that
genomic screening will also create.

Learning more about cancer genomics will
require massive genomic data banks.