Portal de Periódicos da CAPES

Cancer, Our Genes, And The Anxiety Of Risk-Based Medicine

2018; Project HOPE; Volume: 37; Issue: 5 Linguagem: Inglês

10.1377/hlthaff.2018.0344

2694-233X

Siddhartha Mukherjee,

Biotechnology and Related Fields

Narrative MattersPrecision medicine Health AffairsVol. 37, No. 5: Precision Medicine NARRATIVE MATTERSCancer, Our Genes, And The Anxiety Of Risk-Based MedicineSiddhartha Mukherjee AffiliationsSiddhartha Mukherjee ([email protected]) is an assistant professor of medicine at Columbia University, in New York City, and a cancer physician and researcher. He is the author of The Emperor of All Maladies: A Biography of Cancer (Scribner, 2010), which won the 2011 Pulitzer Prize in general nonfiction; The Laws of Medicine: Field Notes from an Uncertain Science (Simon & Schuster, 2015); and The Gene: An Intimate History (Scribner, 2017). The patient's name and certain identifying details were changed by the author to protect her privacy.PUBLISHED:May 2018Free Accesshttps://doi.org/10.1377/hlthaff.2018.0344AboutSectionsView PDFPermissions ShareShare onFacebookTwitterLinked InRedditEmail ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsDownload Exhibits AbstractAdvanced medical technologies make it easier to identify people at risk for cancer, but there are risks involved in oversurveillance, too.TOPICSCancerCancer patientsBreast cancerDiseasesBiopsyCosts and spendingClinicsHealth information technologyProstate cancerIn the summer of 2005 I met a woman, Laura M., whose life had been overturned by cancer. But she was disease free: It was the anxiety of cancer in the future that haunted her.Laura had been diagnosed with a primary tumor in her breast that was small and localized. She had had surgery, radiation, and chemotherapy—a standard treatment protocol—and had then come to see me, an oncologist, to help manage her future care. I suggested doing nothing. Everything about her case suggested a good prognosis: We all agreed that she had likely been cured. But in the wake of her treatment, she became obsessed by the possibility of a relapse. She scoured her family history and discovered a distant aunt who had died of breast cancer at age seventy. Her own mother had died at a young age from a car accident, but Laura became convinced that had her mother lived, she would have been diagnosed with breast cancer.Laura's visits to the clinic were punctuated by her sense of doom. She often came in with sheaves of papers that she had printed from the internet about "occult metastasis," which had been found in patients who, like her, were thought to be at low risk. She repeatedly asked me to confirm that she "had been given the most aggressive chemo regimen that could be given." (She had, in fact, been treated with the appropriate regimen recommended for her case.) We checked her for genetic susceptibilities, such as inherited mutations in the BRCA1 and BRCA2 genes that increase cancer risk, but found none. Nonetheless, she asked if she and her daughter could undergo "the most intensive form of cancer surveillance" to detect early cancers in her body.Laura's story highlights a new anxiety about illness that is permeating our culture. It is the anxiety of being under constant diagnostic surveillance, of inhabiting a state of vigilant watchfulness for illnesses before they can take root in one's body. It is the state, as one patient described it, of "feeling under siege from the future." Emblematic of this anxiety is the concept of a "previvor"—a strange new term invented to describe a person who is a survivor of an illness that she is predisposed to, but has yet to have. For Laura, these states were contiguous: Her survivorship from one breast cancer had turned overnight, it seemed, into "previvorship" for another breast cancer.As I write this, two kinds of technologies are radically altering the landscape of cancer risk and screening. The first involves what we might call "genetic surveillance"—an attempt to quantify an individual's inherited predisposition for cancer (that is, you should be surveyed for the disease because of the higher risk conferred by your genes). The second, in contrast, involves "chemical surveillance"—an attempt to detect chemical markers of incipient cancers in blood (that is, you should be surveyed because there's a sign of early cancer circulating in your blood). The two technologies converge to increase the supply of men and women who are forced to enter the domain of surveillance and screening. Both, in short, encourage men and women without current cancer, but with the prospect of future cancer, to become citizens or permanent residents of what Susan Sontag once described as "the kingdom of the ill."Genetic Screening And Cancer RiskFor decades, perhaps centuries, we've known of families where some form of cancer (usually the same type: breast or pancreatic) is manifest in multiple individuals across multiple generations. Not every family member is affected, but the risk of cancer in such a family clearly lies beyond the average risk of cancer in the population.Until recently, our capacity to identify the culprit genes in such families—or, more actionably, to identify the members of the family who carried the heightened risk—was limited to inherited single-gene mutations. These included mutations in genes such as BRCA1, BRCA2, and MLH1 that, if inherited from parents, increase the likelihood of breast, colon, and other cancers by severalfold over normal individuals. But many human traits, including cancer risk, might not track with single-gene mutations. Take human height as an example. Height is highly heritable—we know that tall parents tend to produce tall children, and shorter parents bear shorter children—yet early attempts to pin down the variation in human height to single-gene variations or mutations revealed only a smattering of candidate genes. (For this essay, I use the terms variation and mutation interchangeably, although there are subtle differences.) Geneticists described this conundrum, famously, as the "missing heritability" of height: We could infer from the pattern of inheritance that height-determining genes must exist in the human genome, but their precise identity and number remained unknown.By similar logic, the inherited risk of cancer might be carried by mutations or variations not in one but in multiple genes, each of which acts together to increase an individual's risk of the disease. In the 1990s some breast cancer patients began to refer to the next, as-yet-unidentified, conglomerate of genes for breast cancer as "BRCA3." That name carried both a sardonic and a hopeful edge. Unlike patients with definite BRCA1 and BRCA2 mutations, patients with potential "BRCA3" mutations remained suspended in an anxious limbo. We could not diagnose a woman with this genetic syndrome yet because we had no idea what these genes might be (a few additional single-gene mutations that increased breast cancer risk were identified in the 2000s, but most patients with breast cancer continued to lack a single-gene explanation). A "BRCA3" patient's experience of her terrifying family history and dread of future disease were just as acute as those of a patient with known cancer-risk mutations, but the genes that precipitated the former patient's fate were hidden from our view. As doctors, we'd acknowledge the risk that these patients carried—with their family histories scarred by breast cancer—but we were unable to offer a more tangible description of their susceptibility to the disease.This state of suspension for "polygenic" (or multigene) diseases is finally being relieved: The combinations of gene mutations responsible for such genetically complex diseases are now being identified by powerful computational technologies. Deep-learning algorithms, in particular, have been unleashed on human genomes. By scanning millions of fully sequenced genomes, these algorithms "learn" to dissect how variations in thousands of genes, each exerting a small effect, might ultimately add up to the heightened risk of an illness—a problem of such mind-boggling permutational complexity that ordinary algorithms had failed to capture it. One machine-learning algorithm has learned to predict human height as the consequence of variations in a thousand-odd genes. (Take a moment to digest this startling fact: Such an algorithm might soon predict your actual height, or the future height of your unborn child, based on your genetic sequence alone.) Another deep-learning program is learning to predict the risk of cardiovascular disease—again, likely the consequence of hundreds of gene mutations or variations. With such advances, it is likely that an algorithm might identify those of us at highest genetic risk for future cancers. (Of course, for many cancers, even ones that run in families, there's still a powerful influence of chance and the environment. A woman with a BRCA1 mutation might increase her risk for breast cancer through certain exposures, by virtue of inheriting other "modifier" genes or by chance alone. These additional variables are not yet part of the deep-learning landscape but could become incorporated into computational algorithms in the future.) This technology, then, could serve as a portal of entry into the world of cancer for potentially millions of men and women who seek to be annotated for future cancer risk and potentially surveyed for cancer.Advancements In Cancer DetectionWhile computers seek out patients who have an inherited susceptibility to cancer, other machines are seeking to identify chemicals that might currently be in our blood or other organs that signal cancer risk. Termed "liquid biopsy," or "liquid surveillance," these methods attempt to discover minuscule amounts of the products shed or spilled by cancer cells—DNA, proteins, and other substances—into the blood or other circulating tissues. Once such trace signs of an incipient cancer are found, the logic runs, cancer will be detected in its earliest stages and can be attacked with more effective therapies. We will scour the body to find ovarian, lung, and prostate cancers, for instance, before these become clinically manifest, thereby enabling better treatments.These liquid biopsies run the risk of overdiagnosing patients, however. What if someone is found to carry a liquid marker for ovarian cancer, say, but that ovarian cancer never takes root in her body? Cancer cells, we now know, can exist in a body, or a site within the body, without becoming manifest as clinical disease or a detectable metastasis. (Most likely, this is because the "soil" of a particular organ does not allow the "seed" of a cancer to sprout.) Or what if some of the markers turn out to overlap with benign diseases (as was the case with earlier liquid surveillance markers, such as the prostate-specific antigen test for prostate cancer), thereby increasing the risk of false positive results?Many patients in my cancer clinic now come to their appointments armed with brochures about liquid biopsies, wondering whether their tumors might have been detected earlier had such biopsies been performed.Nonetheless, enthusiasm for the liquid surveillance of cancer seems to grow exponentially each day (one private company that hopes to advance this technology goes by the name Grail, emblematic of the near-religious fervor with which some advocates describe the power of liquid biopsies). Many patients in my cancer clinic now come to their appointments armed with brochures about liquid biopsies, wondering whether their tumors might have been detected earlier had such biopsies been performed. These technologies represent a second portal of entry into the world of cancer. By identifying men and women who might be bearing the first markers of cancer, these methods increase the pool of those who must be surveyed and further screened for the illness.The Total Institution Of CancerMy aim is to neither exaggerate nor minimize the transformative potential of these technologies, although it's worthwhile emphasizing this at the outset. The capacity to identify humans with an increased genetic risk for cancer, coupled with the possibility of detecting cancer at its earliest stages using a liquid biopsy, might radically change how we prevent, detect, and treat cancer. But my concern is the effect that such surveillance might have on our bodies and societies. In the 1950s the sociologist Erving Goffman wrote a remarkable article about the concept of a "total institution," an idea that he expanded in subsequent work. "A total institution," Goffman wrote in his 1961 book Asylums, is one "where a great number of similarly situated people, cut off from the wider community for a considerable time, together lead an enclosed, formally administered round of life." Total institutions, such as mental hospitals, prisons, and even boarding schools, have rituals of entry and exit. They inculcate belonging. They invent their own vocabulary and codes of behavior; they have an internal logic, impenetrable to others. They encourage surveillance and create anxiety: Members are united by a common sense of purpose, by the feeling of being chosen or marked. Those who are expelled may feel a sense of betrayal, while those who remain can be consumed by the guilt of survivorship.Cancer, too, runs the risk of becoming a "total institution." A patient, once diagnosed, may be whisked away into a cancer ward, dressed in a patient's smock—"a tragicomically cruel costume, no less blighting than a prisoner's jumpsuit," as I wrote in The Emperor of All Maladies—and stripped of his identity. When I once asked a woman with a rare sarcoma about her life outside the hospital, she observed, "I am in the hospital even when I am outside the hospital."In this new era of cancer treatment, I wonder if we are unwittingly, but insidiously, intensifying the totality of the "cancer institution" for patients. For people like Laura M., cancer has certainly become a total institution—or a "cancer world," as some patients call it. They are in either treatment, remission, surveillance, maintenance, or resurveillance. Mavens of early detection are also working on deep-learning algorithms that will pick up cancerous lesions on patients' imaging results and classify them as malignant, using criteria that seem to defy even the most acute human eye. In an April 2017 article in the New Yorker, I wrote about one of the pioneers of this idea, the German computer scientist Sebastian Thrun. Thrun imagines a world in which even the daily instruments of our normal lives are morphed into weapons of diagnostic surveillance—a bathtub that scans your body to detect abnormal masses that might require investigation; a mirror that could check your body for precancerous moles; a computer program that (with your consent) would scour your Instagram or Facebook page while you slept at night, evaluating changes in your photographs that might signal signs of cancer.Then there's the question of treatment and cost. If an additional tumor—clinically undetectable, but discovered by these novel methods—were detected in Laura M.'s case and the primary lesion removed, what criteria would we use to determine whether we should use some form of adjuvant (or extra) medicine, such as cell-killing chemotherapy or targeted therapy, after the initial surgical removal, as is often done for most cancers? The costs of such surveillance and treatment—an astronomical amount if every human had to be genetically annotated, subjected to surveillance, and treated if a tumor was found—would overwhelm current projections of medical costs (although in the most optimistic scenario, the benefits would also be amplified in lives saved via early diagnosis). Thorny issues of overdiagnosis and overtreatment would have to be addressed. We would have to devise careful guidelines about when not to act and whom not to treat.For Laura M., the answer to each of these questions carries immense consequences. She has entered a strange new world, one of constant diagnostic surveillance; of dealing with the anxiety of relapse and maintenance; of that peculiar desolation of the shuttle from clinical trial to clinical trial, and from hospital to hospital, as she tries to keep one step ahead in this chess game against cancer; and of watching doctors pit their will, wit, and imagination against a formidable enemy that keeps changing its shape. This world has created its own internal vocabulary. A "haircut party" is a celebration thrown in honor of a person about to enter the cancer world (as a sign of solidarity, even if the patient is spared hair-loss inducing chemo). "No Exit chemo," as a patient of mine put it, describes the fact that a unique personalized chemo regimen for a patient produces unique toxicities, a phrase borrowed from the Jean-Paul Sartre novel in which every human being is assigned his or her own personal hell."A world in which cancer is normalized as a manageable chronic condition would be a wonderful thing," the medical historian Steven Shapin wrote in a 2010 review of The Emperor of All Maladies. "But a risk-factor world in which we all think of ourselves as precancerous would not," he continued. "It might decrease the incidence of some forms of malignancy while hugely increasing the numbers of healthy people under medical treatment. It would be a strange victory in which the price to be paid for checking the spread of cancer through the body is its uncontrolled spread through the culture."Laura M. has not suffered from a relapse of breast cancer. Nor, fortunately, has she had a new cancer anywhere in her body. But the "strange victory" over her body has not spared her mind.To date, Laura M. has not suffered from a relapse of breast cancer. Nor, fortunately, has she had a new cancer anywhere in her body. But the "strange victory" over her body has not spared her mind. She remains haunted by the future prospect of illness. When Susan Sontag wrote of a passport between the kingdom of the well and the kingdom of the ill, she imagined a bidirectional passage: men and women might pass into illness, but some would return to wellness. In inventing cancer's new surveillance culture, I fear that we have closed the borders of the kingdoms. I fear that we now possess just one-way passports into the realm of illness. Loading Comments... Please enable JavaScript to view the comments powered by Disqus. DetailsExhibitsReferencesRelated Article MetricsCitations: Crossref 2 History Published online 7 May 2018 Information© 2018 Project HOPE—The People-to-People Health Foundation, Inc.PDF downloadCited byPersonalized Screening and Prevention Based on Genetic Risk of Breast Cancer19 March 2022 | Current Breast Cancer Reports, Vol. 14, No. 2Sequencing Newborns: A Call for Nuanced Use of Genomic Technologies14 August 2018 | Hastings Center Report, Vol. 48