When Bryce Olson tells his prostate cancer story, he pinpoints the conversation he had with his oncologist six years ago as the moment his life changed. That’s when Olson, then 44, his body pumping with adrenaline and his hip throbbing with new pain from a metastasized tumor, uttered these words: “I want to get my cancer sequenced.”
Wearing a slickly designed black T-shirt with the words “Sequence Me” in block heavy-metal-style lettering, Olson preaches to audiences at health care conferences that they should take back control of their health from paternalistic providers and demand better cancer care. On his website, he offers patients a “battle card” with scripted comebacks to doctors’ resistance.
And if your doctor refuses to order genetic testing? “Well, then you need a new doctor!” he exclaims to often enthusiastic applause.
But in 2014, his own doctor at the Oregon Health & Science University medical center in Portland, after hearing Olson’s request, had merely asked, “Why?” It didn’t make sense, the doctor had said, especially if it turned out there were no drugs available for the DNA mutations the sequencing technology would find.
Despite the handful of so-called targeted cancer therapies that were matched to patients’ genetic makeup, there wasn’t one for prostate cancer that had been approved by the US Food and Drug Administration.
Olson was already aware of that fact. The conversation was life-changing, because it marked the moment he was done with the American medical system—or at least what’s known as “standard of care” for cancer patients.
By that time, Olson had already undergone surgery to remove his prostate gland and seminal vesicles. Soon after, he learned the cancer had spread to his bones—creating lesions the size of quarters—and was incurable. That was the news that made him sob hysterically inside his car in the hospital parking garage while blaring the radio and repeatedly screaming “Fuck!” He’d researched the statistic that the median five-year survival rate for Stage 4 prostate cancer patients was 28 percent.
Olson also had just finished six months of chemotherapy. Despite hearing about a “golden age” of cancer treatment, he was prescribed the drug docetaxel, which had been patented in 1986 and was routinely given to patients with breast, lung, prostate, stomach, and head and neck cancers. The treatment, sold under the brand name Taxotere, was so toxic it had earned the name “Taxo-Terrible,” because of its dreaded side effects. Olson got them all: zero energy, hair loss, bone pain, mouth sores, constipation, and a numb sensation in his feet that felt like needles were poking him and made it hard to keep his balance.
The mother of his 6-year-old daughter’s best friend had been so worried about him she asked the school district to move up the annual father-daughter dance from spring to fall to make sure he could attend. During those months, he’d tried to spend as much time with his kid as possible.
At home, he’d play guitar while she drew, or he streamed his favorite ’80s movies, such as Breakfast Club and Pretty in Pink. Even though she was too young to understand the references, he wanted to make sure she didn’t miss out on what he considered an important part of his childhood. He also gave her pep talks about avoiding drugs and only dating boys who treated her right—in case he didn’t get another chance later.
Olson was grateful she didn’t grasp how sick he really was, so he wrote and recorded songs for her to listen to when she was older. His favorite lyric from his song “My Girl”: You won’t see me on the outside. But I’ll be there on the inside.
Some days he was angry at the world, and other days he was overwhelmed with sadness as he sat on the couch in front of the window and watched people walking by, knowing they were planning their weekends and summer vacations. He wondered about mundane details: Should he buy new winter clothes if he would only end up wearing them for a season? Brushing his teeth felt like a ridiculous waste of time.
Despite surviving Taxo-Terrible, Olson figured that the sharp pain in his hip meant the cancer had spread there too. He wasn’t interested in the next recommended treatment. According to the National Comprehensive Cancer Network guidelines that shape the decisions of the nation’s oncologists, his options at the time would have been another old-school hormone therapy or an immunotherapy that was approved in 2010 and extended patients’ lives by a median of four months.
“I want to know what’s driving my cancer at a molecular level,” he had told the doctor.
In 2014, even though Oregon Health & Science was one of the few medical centers in the country that had set up an in-house DNA sequencing lab, Olson knew his request was unusual. But as an IT marketing executive with the supercomputing giant Intel for 15 years, he also knew more than most people about the genetic science known as genomics. After being diagnosed with cancer, he had asked to join his company’s newly formed health care team, which was working with several institutions to create tools to process vast amounts of genomic data.
Olson hoped he could use his DNA results to inquire about an off-label use of an existing cancer drug or get on a clinical trial that would give him a fighting chance of survival. Still, his doctor said it was too risky to bet on an unproven drug. “I may not recommend it at this time,” he had said.
But he agreed to put in the order.
Today, we truly are in a golden age of cancer treatment.
In 2019 there were 16.9 million cancer survivors in the United States—a number that’s expected to grow to 21.7 million in 2029. During the past couple of years, the FDA has approved an unprecedented number of cancer drugs, including two new ones in May for metastatic prostate cancer patients with certain genetic mutations. There is now a once-unimaginable number of targeted therapy drugs for more than 30 kinds of cancer—and some 1,100 new cancer drugs and vaccines are in the development pipeline. Of the cancer clinical trials conducted last year, 61 percent involved the use of biomarkers, compared to 18 percent in 2000.
There’s even more government funding: This year, the National Cancer Institute received a $297 million boost to its annual $6.4 billion budget, which paid for more than 125 additional research projects.
These milestones reflect the vision set forth by President Barack Obama when he announced the Precision Medicine Initiative during the State of the Union Address in January 2015—a little more than a month after Oregon Health & Science pathologist Christopher Corless analyzed Olson’s walnut-sized prostate gland that had been excised and embedded in wax. Corless cut the tissue into thin wafers using a device that looked like a deli pastrami slicer.
After the tissue was stained to identify the cancer under a microscope, the malignant areas were placed in a test tube and mixed with an enzyme to free its DNA secrets.
The resulting clear fluid was dropped into a sequencing machine that for 24 hours spit out strings of 150 letters and stitched them back together to examine the 37 genes most associated with prostate cancer. (Today many labs analyze anywhere from 300 to 600 genes across multiple kinds of cancer.) The machine compared the alterations in Olson’s tumor against a “reference set,” which is made up of population averages of healthy DNA, to see which ones were abnormal. Then Corless had to consult the literature to figure out which ones might be making Olson sick.
“Doctors have always recognized that every patient is unique, and doctors have always tried to tailor their treatments as best they can to individuals. You can match a blood transfusion to a blood type—that was an important discovery,” Obama had orated in his booming voice. “What if matching a cancer cure to our genetic code was just as easy, just as standard?”
America had entered a new age of personalized medicine. A year later, in 2016, Congress passed the landmark 21st Century Cures Act that earmarked $1.8 billion in funding for then vice president Joe Biden’s Cancer Moonshot task force over seven years to accelerate cancer progress. Among the blue-ribbon panel report’s goals were compiling patients’ genetic tumor data, developing new technologies to characterize tumors and test therapies, and creating a “national ecosystem” to promote sharing and analyzing of cancer data.
In the meantime, the nation’s academic cancer centers staffed up labs, trained clinicians, and created “molecular tumor boards” that met regularly to review patients’ genetic results. Private testing providers, such as Foundation Medicine, Tempus, Guardant Health, and Caris Life Sciences, also filled in the gaps by offering sequencing of solid tumors and, most recently, blood-based “liquid biopsies” that can detect the tiny fragments of DNA that a tumor or blood cancer sheds. Their reports come with a list of FDA-approved and off-label drugs and relevant patient clinical trials.
Yet as Olson passionately points out on the conference circuit, getting oncologists to order DNA testing in the first place is still a challenge. A June report by the Personalized Medicine Coalition found that in 2019, medically necessary testing was “inconsistently utilized” across the United States, despite the fact that insurers increasingly covered it for metastatic patients.
“What’s needed is a lot of culture change, especially with doctors who don’t regularly use it in practice,” says Monica Mita, codirector of the experimental therapeutics program at the Samuel Oschin Comprehensive Cancer Institute at Cedars-Sinai Medical Center in Los Angeles, who recruits patients for clinical trials. “Some don’t have the time to do genomic profiling for their patients or aren’t comfortable commenting on the results or making recommendations of treatments they may not be familiar with.”
Other surveys, including one about oncologists who work in community practices, where 85 percent of US cancer patients are treated, have pointed out that doctors are also worried about the cost to patients—both in surprise medical bills and dashed hopes. What do they tell patients when the shiniest new sequencing technology doesn’t lead to a miraculous cure?
As with any new technology, there are roadblocks to adoption: Clinicians need to master a steep learning curve that includes keeping track of new tests, drug approvals, clinical trials, and constantly changing treatment guidelines.
That’s not to mention the exploding amount of new genetic information these tests are pumping out, which researchers are racing to interpret and make sense of. They’re encountering what are called “variants of unknown significance,” and oncologists are often stuck trying to explain to patients why they shouldn’t be worried about those genetic mutations now—but there may come a day when that information will mean something.
Yet the part of the story that has gobsmacked everyone—policymakers, researchers, doctors, tech executives, privacy and patient advocates—is a problem that from the outset seemed much simpler: What do we do with all the data about the variants that do matter, and how do we make it useful for patients?
In other words, how do we take all the collective genomic data that has been generated thus far and figure out why certain people with certain genetic mutations and cancer status respond to certain drugs?
Biden imagined such a world in the Cancer Moonshot plan that he submitted in 2016 to President Obama, writing: “We sought to better understand and break down the silos and stovepipes that prevent sharing of information and impede advances in cancer research and treatment, while building a focused and coordinated effort at home and abroad.” (He has lamented that while his son Beau was being treated for brain cancer, the family couldn’t get health records sent between Walter Reed National Military Medical Center in Bethesda, Maryland, and the University of Texas MD Anderson Cancer Center in Houston, because the systems weren’t compatible.)
Yet that effort so far has been profoundly disappointing. There are many challenges, such as figuring out how to organize, standardize, and store vast amounts of genomic, lab, and clinical data—and get them in and out of the nation’s exasperating sea of electronic health records while still protecting patients’ privacy.
But the bigger barrier is more daunting: We live in a country in which private health institutions don’t share patient data. “When competitive information is involved, sharing is limited,” says Atul Butte, director of the Bakar Computational Health Sciences Institute at the University of California, San Francisco. “American Airlines doesn’t share with United or Delta.”
Adds biomedical engineer Steven Salzberg, director of the Center for Computational Biology at Johns Hopkins University: “Researchers’ fame is in their own data. Their motivation is to advance their own careers. They need to publish papers and get grant funding,” he says. “People will share data, but they need outside pressure to do it.”
However, if you’re sick with cancer—and you’re holding a brand-new DNA report that reveals the exact code for your kind of cancer—you’re going to want to know which treatments have worked for other patients like you.
What if their doctors know something that yours doesn’t?
At the beginning of 2015, Olson learned he had an alteration in the PI3K signaling pathway, an important regulator of cell growth. Plus, he had a mutation in his PTEN tumor suppressor gene that caused his cancer to become hyperactive. His doctor helped him find a clinical trial for a drug known as a PI3K inhibitor that was developed by Verastem Oncology, a biopharmaceutical company based in Needham, Massachusetts.
But when Olson called Monica Mita’s group at Cedars-Sinai about getting on the trial, he was told it was full—that is, until he uttered these words: “I have sequencing data,” he said. “I’m a perfect match.”
Olson traveled from Portland to Los Angeles every few weeks to receive the drug and undergo monitoring. Within six months, his prostate-specific antigen, the protein secreted by cancer cells that oncologists use to measure a patient’s cancer, went down to zero, his bone lesions mostly disappeared, and Olson started to feel like his old self again.
The drug known as VS-5584 was working.
Every six weeks or so, Olson’s stomach was in knots as doctors sampled his blood and scanned his body—a phenomenon known among cancer patients as “scan-xiety.” But he always received the news that the cancer was contained. “As long as you’re doing great,” he recalls his doctor saying, “you could be on this indefinitely.”
Cancer survivors often talk about their disease as a gift, because of its power to make people focus on what’s important in their lives. When time is precious, you don’t stick around in bad relationships or soul-sucking jobs. When Olson had imagined what his obituary would say, he figured it would include that he was a great dad, son, and friend. He liked to surf and ski and had a tight group of friends who loved to get dressed up as ’80s rockers and sing karaoke around Portland. He had figured he’d stay at Intel until he could retire, then split his time between Costa Rica and the Oregon coast. Olson had an enviable life. He just didn’t think it had a lot of meaning.
With genomics, he suddenly felt a calling to spread the word about how it had saved his life. A longtime singer-songwriter, Olson recruited a half-dozen local artists, including Jenny Conlee of the Decemberists and Martha Davis of the Motels, to put together an album to spread awareness about DNA sequencing called FACTS (Fighting Advanced Cancer Through Songs). Wearing his “Sequence Me” shirt, Olson launched it in March 2017 at SXSW in Austin, Texas.
During this time, Olson also came to the slow, painful realization that his marriage had run its course. Married at 22 to the college sweetheart he had met when he was a sophomore and she was a senior, they’d settled into a pattern: Olson slept in the basement, and they had separate friends and activities. “When you think you’re going to live until 80, you figure you can suck it up until your kid goes to college,” he says. But he got a burst of clarity when he realized he didn’t want his daughter to think this was what a healthy marriage looked like. Olson was suddenly more scared of dying in an unfulfilling relationship than dying alone.
A year into his trial, however, Cedar-Sinai’s Mita told him it was ending. The FDA was worried that the drug, whose safety was still being tested in a Phase I clinical trial, was too toxic to the heart. Then came the other bombshell: Out of 75 patients across four sites nationwide and one in the United Kingdom, Bryce Olson was the rare success story. Under the FDA’s compassionate use program, Olson received leftover drugs from the investigators at the other sites: a nine-month supply.
Olson started to panic. He had no idea what he would do once the drug ran out. He called Verastem, pleading, “You can’t just cut me off,” and asked if he could get the formula so he could find his own manufacturer. But on Christmas Eve, Olson learned the company had returned the patent to the inventor. He was warned that he might develop heart complications too.
But by March 2017—the same month he was due to run out of his supply of VS-5584—Olson’s stomach fell for another reason. When he logged on to a patient portal to look at his most recent PSA numbers, he saw they were up.
After two full years in remission, the cancer was back.
When Olson’s doctors had suggested the Verastem trial, the decision to pursue a drug that targeted the pathway linked to his cancer seemed obvious.
Yet here’s the moment when the glow of the genomics revolution starts to dim: What happens when a drug stops working? How does all that genetic information—a windfall patients could only have dreamed of a few years earlier—continue to be helpful, especially when the DNA of cancer is more complicated than anyone could have realized?
“We know that advanced cancer changes quickly,” says Brian Druker, director of the Oregon Health & Science’s Knight Cancer Institute, whose research in the 1990s led to the discovery of Gleevec, a breakthrough targeted treatment for leukemia. “Targeting one genetic abnormality can get a great response, but then it changes and becomes resistant. Other pathways can take over for the original mutation and drive the cancer.”
This is why Nicholas Schork, deputy director of quantitative science at the Translational Genomics Research Institute in Phoenix, says that cancer research could benefit from what he calls a “monster database,” with algorithms that could show nuance and patterns.
“If you’re an oncologist and you don’t know what drug to give a patient, if you could match your patient’s sequenced tumors to others and know how they’d been treated and how they responded, you would be more likely to find a drug that works,” Schork says.
Even better, what if that database also included other important information, such as what other treatments patients had failed, the stage and characteristics of their cancer, and how it had mutated and developed resistance?
Schork envisions two scenarios. One: Private companies could put together such a database and then charge an access fee. The question, however, is whether it would be fair for patients who couldn’t afford it. “It could compromise their ability to be treated for cancer. It would compromise their life,” he said. “Are you creating a situation in which only the haves—as opposed to the have-nots—are benefiting?”
Or, two, such information could be perceived as a public good that everyone would have access to, perhaps funded by taxpayer money and run by a public agency like the National Institutes of Health. “This is a debate the medical community needs to have,” Schork says. “This is something that we as a society will have to deal with.”
We’ve been here before. In 2013 the US Supreme Court made a statement about the ownership of biological information by ruling that human genes couldn’t be patented. In the landmark unanimous decision, the justices said the company Myriad Genetics’ claims that mutations to the BRCA1 and BRCA2 genes, which are linked to increased risk of breast, ovarian, and prostate cancer, were “invalid because they covered products of nature.”
There are other government precedents that genomic information should be perceived as a societal good. In 2014 the NIH issued its Genomic Data Sharing Policy, which requires any agency-funded scientist to release de-identified human data into an open research repository.
“These are top-down efforts to affect change so that data sharing and open access are baked into the ethos of funding,” says oncologist Eliezer Van Allen, who runs the metastatic prostate cancer project of the Broad Institute’s “Count Me In” initiative. The program has collected blood and tissue samples directly from more than 8,000 patients for five different types of cancer in conjunction with the Dana-Farber Cancer Institute in Boston. The samples are sequenced in-house, and then the aggregated data is distributed through NIH channels.
And in the NCI’s Molecular Analysis for Therapy Choice project that was launched in 2015, trial investigators are studying the effectiveness of drugs against targeted genetic changes in nearly 40 treatment arms.
Another way the government forces data sharing is through its 2007 requirement that the majority of drug and medical device investigators register clinical trials and report their findings to the federal database ClinicalTrials.gov, which was created a decade earlier.
Yet a Science investigation published earlier this year revealed that investigators routinely flout the requirement and go unpunished, despite the NIH and FDA’s attempt in 2017 to increase enforcement.
After examining more than 4,700 trials, the magazine found that out of 184 sponsor organizations with at least five trials due last September, “30 companies, universities, or medical centers never met a single deadline … [and] failed to report any results for 67 percent of their trials.”
“Compliance isn’t as good as it should be,” says Deborah Zarin, who headed ClinicalTrials.gov between 2005 and 2018 and now studies clinical trial design and reporting at Brigham and Women’s Hospital in Boston. This August, the FDA threatened a new round of legal action and potential fines.
In the meantime, many medical journals are resorting to another kind of pressure to compel data transparency—by making it a factor in whether studies get published in their pages. Starting in 2019, the International Committee of Medical Journal Editors requires investigators to disclose their data-sharing intentions in advance when they register a trial. In addition to providing summary results, they have to include whether they plan to reveal de-identified participant data to show how individuals respond to an intervention.
The potential benefit is that such information about the people in studies might be helpful to doctors making treatment decisions, Zarin says, calling the editors’ position an “important trend” that increasingly puts data sharing on researchers’ radar.
Although sharing this additional layer of data isn’t mandatory, the committee writes, it could easily be seen as an expectation, especially when you consider this sentence: “Investigators should be aware that editors may take into consideration data sharing statements when making editorial decisions.”
Still, despite efforts to encourage more transparency, many investigators disclose shockingly little information, especially when it comes to failures, which account for about 90 percent of trials. According to FDA guidelines, researchers are required to provide “any available data” for trials that are stopped midway. Yet anyone interested in learning more about the fate of Verastem’s VS-5584 would glean little from the company’s report on ClinicalTrials.gov. The recruitment status simply reads: “Terminated (Lack of recruitment and the company’s decision to-deprioritize 5584 development.)” There are no results posted.
Such a lack of information is particularly frustrating for cancer patients researching new therapies or scientists trying to develop new ones, says Andrea Miyahira, director of global research and scientific communications at the Prostate Cancer Foundation in Santa Monica, California.
“Right now there are a number of trials testing patients for genetic mutations. There are also many patients being treated with experimental therapies, but the reporting is inconsistent or nonexistent,” she says. “It’s also common to see case reports in the literature in which patients exhibit exceptional responses but with little other data on the patients who didn’t respond.”
Or they don’t include patients like Olson who thrived at first and then stopped responding altogether.
After the PI3K inhibitor made by Verastem stopped working in 2017, Olson went back to Monica Mita at Cedars-Sinai to ask about getting on another trial. This time, she found a Phase I trial for another kind of PI3K inhibitor that was made by the pharmaceutical company Eli Lilly. The closest location, however, was in Oklahoma City. So Olson began flying there every two weeks from Portland.
Olson knew he was lucky. As an Intel executive, he could afford to hop on planes. Fewer than 5 percent of adults with cancer are enrolled in clinical trials, and the cost of travel is a big reason. So is having to take time off work and find childcare.
But within two months, Olson dropped out. His PSA had doubled. He worried his cancer had changed and stopped responding to PI3K inhibitors.
The news got worse: He had a new lesion on his T5 vertebra. He couldn’t stop thinking about a recent Duke study that found the median survival time for prostate cancer patients who had cancer in their bones and had already failed chemotherapy was 21 months.
By then, Olson had stopped widely sharing updates, especially with his parents. Ever since his kid brother was fatally stabbed during a home invasion in 2011, Olson felt a responsibility not to compound their sadness. They had been through enough heartbreak.
So he asked a pathologist to scrape DNA from the lesion, but the sequencing results revealed the same handful of mutations, and there weren’t any good new drugs or trials to try.
At the same time, Olson was getting more attention for his Sequence Me project. He was invited to speak at the cancer drug manufacturer Roche’s annual meeting and conferences, including the Nantucket Project, the National Society of Genetic Counselors, and the Consumer Electronics Show in Las Vegas. He was also written about in Forbes and STAT.
The irony was not lost on him. Olson had become a genomics evangelist. Yet the latest crop of treatments had stopped helping him.
What frustrated him most was that he knew enough about cancer research by then to know that nothing was conclusive. There were several kinds and classes of breakthrough drugs that simply hadn’t been tested on his kind of cancer yet. There were new combos, doses, targets, and mechanisms of action. Plus some of the newest genomics discoveries had little to do with naming a tumor’s DNA, but rather measured other biomarkers such as the number of tumor mutations and a cell's ability to repair itself. Other research focused on a cancer’s “microenvironment” and how there could be different ones throughout the body.
Yes, cancer treatment was part luck and part art. Oncologists often likened it to whack-a-mole: After one method ducked out of sight, you kept hitting whatever showed up next with whatever hammer science had to offer that particular month or year.
If you knew the next thing to try.
“Someone at a place like Memorial Sloan Kettering who looks like me genomically is taking a drug combo that is working magically,” he says. “But I’ll never know what that is because we don’t share it.”
So Olson set out to try to create his own database. He reached out to cancer genomics researcher Pablo Tamayo at the University of California, San Diego, to see if he would take on a new research project comparing the molecular fingerprints of Olson’s cancer against others. Tamayo and two colleagues found a couple of public data sets of more than 500 advanced prostate cancer patients and began to hunt.
They homed in on eight patients who seemed like good matches. “But we didn’t know if any responded to any of the treatments Bryce was interested in,” said Tamayo, director of the UCSD Center for Cancer Target Discovery and Development. “The system isn’t set up to do that.”
Then they asked the Broad Institute’s Van Allen, who’s also an associate professor focusing on computational cancer genomics at Dana-Farber Cancer Institute in Boston, to see if he could fill in the gaps. Van Allen had to call doctors he knew personally to see the de-identified records of those eight patients. But none had received any of the experimental therapies Olson wanted to know about, and by the time they got the data, half had already died.
By this time, his new oncologist, Rana McKay of UCSD Moores Cancer Center, was becoming concerned: Olson’s PSA was rising while he was chasing futile research results without a drug in his system.
He’d met McKay in 2017 through the colleague of a colleague in the cancer research community. The young doctor had a reputation for being a leader in data-driven precision medicine, and she was particularly interested in why prostate cancer became resistant to drugs. Olson immediately loved her warm demeanor and willingness to be creative in her treatment approaches. He called her his “oncology soul mate.”
McKay had decided to become a cancer doctor in high school after watching her mother suffer from debilitating nausea while undergoing chemo for breast cancer in her thirties; McKay decided to specialize in prostate cancer because breast cancer was “too close to home,” she said. (Her mom is still in remission.)
But the need for fresh thinking was just as urgent, given that one in nine men will be diagnosed with it during their lifetimes. Although most men develop the disease at an older age and ultimately survive (or die from other causes first), Olson’s kind, known as castration-resistant because it grows even when testosterone levels are low, accounted for 10 to 20 percent of the nearly 192,000 estimated new cases in 2020 and was particularly lethal.
McKay was deeply disturbed that there had been little progress in the field and was inspired by a seminal 2015 paper in the journal Cell that found 89 percent of patients with metastatic castration-resistant prostate cancer had a “clinically actionable aberration”—that could be targeted by an available drug—and that sequencing could “impact treatment decisions in significant numbers of affected individuals.”
“There are a lot of nuances to prostate cancer. It’s not cookie-cutter at all,” says McKay. “We’re learning more about genomic subsets and how these tumors can be vulnerable to specific therapies. But if you never test for them, you’ll never find them.”
At the time, she didn’t have any creative approaches to try with Olson. So she made him a deal: They would continue to mine genomics, but he would focus on slowing the growth of the cancer by taking a standard-of-care hormonal treatment in addition to steroids.
That approach worked for 10 months. Then in April last year, the cancer spread to his ribs. McKay sent off bone and blood samples to the private testing companies Foundation Medicine and Tempus. While they waited for results, McKay called him excitedly with the news that Olson qualified for a new clinical trial at UCSD. It had nothing to do with genomics. Rather, it was for an immunotherapy and targeted-treatment combination that had worked wonders on renal cancer carcinoma patients and now was being tested on patients with his exact kind of prostate cancer.
It would be Olson’s seventh line of treatment in six years.
When people indulge in lofty debates about whether genomics data should be widely shared, they often forget one critical detail: how hard it is to actually share it.
Raw sequencing data is massive, unwieldy, and often does not come in an easily readable format. And despite a $48 billion federal mandate begun in 2009 as part of the 21st Century Cures Act that would make all our health records available electronically and compatible through a user-friendly application programming interface, our DNA readouts won’t likely be part of the package anytime soon.
In March, the US Department of Health and Human Services finalized a controversial rule that would require health record vendors to make patients’ information available using smartphone applications at no cost. The big idea is that patients would be in charge of their own medical records and drive their own data revolution. They could authorize third-party apps to crunch it into something useful—perhaps sending patients personalized treatment recommendations as well as GPS location pins of available clinical trials.
Apple currently offers iPhone users the ability to view their own health records—and share them with app developers through its Health app. It now has more than 500 participating institutions, “We think there will be a vibrant economy of folks that patients will trust with their data,” says Donald Rucker, who heads the Office of the National Coordinator for Health Information Technology. “The patients will direct providers to give developers their data through the app. There will eventually be disease-specific ones, and the apps will compete with each other.”
Yet the initial focus would be on basic information, such as lab results or doctor’s comments. “Genetic sequencing might be at the far end,” Rucker says.
In the meantime, private enterprise has stepped forward to try to create the “monster databases” Nicholas Schork had described.
Foundation Medicine, a commercial sequencing lab based in Cambridge, Massachusetts, that was bought by Roche in 2018 for $2.4 billion, is mining its more than 450,000 patient profiles—as well as more than 64,000 electronic health records from cancer data company Flatiron Health (also bought by Roche for $1.9 billion in 2018).
A 35-person team combs through medical journals, NIH and NCI data sets, and other research in the public domain to connect the dots between genomic signatures and treatment outcomes to generate insights for the reports it provides to oncologists and patients. The company is also working on penetrating the silos of private institutions. In August it announced a data-sharing agreement with OneOncology, a group of nearly 170 community practices across the country.
Yet aggregating that data is a monumental task. “It requires a lot of cleaning and structuring,” says Foundation CEO Cindy Perettie. “One group might call something ‘nausea,’ while another will call it ‘vomiting,’ for example.”
Meanwhile, Chicago-based Tempus—created in 2015 by Groupon founder Eric Lefkofsky after he was shocked by the standard-of-care options and lack of good data when his wife was diagnosed with breast cancer—uses optical character recognition and natural language processing to pull information from patient records. It claims to have access to the data of one-third of US cancer patients through contracts with hundreds of academic centers and practices.
As part of the deal, oncologists order their tests, and Tempus offers access to its database through an analytical tool cleverly named Cohort that contains de-identified patient information and treatment outcomes. “We give back all of the data that we sequence,” says Lefkofsky, CEO of Tempus, which has raised $620 million from investors and is now valued at $5 billion. (It has since expanded into depression and Covid-19.)
The goal is to empower doctors with the most current data. “There’s still a decision point that oncologists can make if a line of therapy doesn’t work,” Lefkofsky says. “They can base some of these decisions on the data.”
Despite Schork’s concerns that poor patients would be shut out of such technology, both Perettie and Lefkofsky say their products, which range from $3,800 to $5,800 retail, are often covered by insurance, and they offer financial assistance programs that generally cost patients less than $100.
Just like the direct-to-consumer genetic testing company 23andMe, they make most of their money through deals with biotech and pharmaceutical companies. For example, they help fill clinical trials that need participants with specific genomic profiles or develop what are called companion diagnostics to match patients with certain drugs. Earlier this month, the FDA approved Foundation’s tissue test to identify metastatic castration-resistant prostate cancer patients with certain genetic mutations who might be candidates for the targeted therapy olaparib, which is sold under the brand name Lynparza by the pharmaceutical company AstraZeneca.
Giving patients access to genomic profiling isn’t just a question of fairness, says Megan Roberts, director of Implementation Science in Precision Health and Society at the University of North Carolina Eshelman School of Pharmacy. It’s vitally important to include as many data points as possible to create a more useful database for everyone.
That’s the rationale behind the NIH’s historic “All of Us” program that was launched in 2018 to collect the genomic and health information from 1 million diverse US residents and improve the unsettling fact that the vast majority of genomic information comes from people of European or Asian descent. “We know there are racial or socioeconomic inequities as to who gets testing,” Roberts says. “That means these databases include more of the privileged few who were lucky to have access to precision medicine. We need to be careful not to widen disparities in cancer care outcomes.”
Given the size and speed of these shifts in the testing market, critics are left asking: Is gathering, organizing, and distributing our collective biological information an appropriate job for technology companies? Lefkofsky defends his company’s role. “It has to be done by private enterprise. The cost of doing it on the clinical side is too big for the public to afford,” he says. “We want to be able to offer this information at scale.”
However, bioethicist Michelle McGowan at Cincinnati Children’s Hospital Medical Center in Ohio, who studies health care equity, believes the federal government is better positioned to act as an honest broker. The reason is simple, she says: “It doesn’t have the same financial skin in the game as private companies do.”
She suggests the government—perhaps pushed by vocal patient advocacy groups—can do more to make private health care institutions more transparent. The fertility industry offers a good example, she says. Since 1992, clinics routinely report their in vitro fertilization success rates to the CDC, which are made public on its website. “This was framed as a consumer protection model. Although the reporting is voluntary, it’s become an expectation. One could imagine this happening in the cancer space too,” McGowan says. Imagine your doctor posting your cancer status, genomic profile, and treatment-regimen outcome.
Yet there’s another force at work that might be more compelling than government regulation. It’s called peer pressure, and the catalyst was Covid-19. Doctors needed to quickly know how the coronavirus affected cancer patients and were forced to experiment with new models.
When the Covid-19 & Cancer Consortium was launched in March, researchers collected data using what’s known as a federated approach: giving access to specific data from one’s institution while safeguarding the integrity of its databases and the privacy of patients. Now more than 120 cancer centers have contributed de-identified information for nearly 6,000 patients. “There are ways to keep things secure and allow them to be shared, especially when patients have consented to it,” says McKay, who’s a member. “It was a great feat of academic collaboration.”
The government is currently funding a project that’s testing a similar idea among five children’s hospitals. Kenneth Mandl, director of the Computational Health Informatics Program at Boston Children’s Hospital, received a $9 million grant to create infrastructure so that institutions can share protected data. (His team just finished building the federally regulated interface of an “app store for health” model to run on health IT systems.)
The first task was to standardize the consent process, electronic health records, and genomic data at all institutions. The data stays local, yet researchers at other hospitals can easily query it. “The idea is to make it a turnkey experience,” Mandl says. “If the model works, we could get it to a large scale, and a large proportion of the population could participate.”
In January, Olson turned 50—a once unimaginable milestone. He and 16 friends celebrated by dressing up as ’80s hair-band rockers and renting a stretch limo that took them around Portland to their favorite dive bars and a Guns N’ Roses tribute-band show.
His body is no longer the same after six years of treatment, which includes a constant low dose of hormone therapy that shuts down his testosterone. Olson still has his boyish Nordic good looks, but he’s lost some of his physical energy and muscle mass. Before Olson got sick, he used to be able to grow a respectable Tom Selleck–style ’stache, but now he has little facial or body hair. In a strange silver lining of cancer, he has more hair on his head now than he did at 43.
And in the biggest shock of all, Olson swears that sex is better—a point he wants all prostate cancer survivors to know, because it’s rarely discussed. He wants to push back at our culture’s obsession with testosterone—and marketing of testosterone replacement therapy—and the notion that masculinity is defined by a ripped physique and robust erection.
In 2014, when his surgeon removed half of the nerve structures near his prostate gland known as neurovascular bundles, Olson had been terrified that his sex life was over.
He worried that his lack of testosterone would gut his desire. So after meeting with a sex counselor who reassured him he could still have satisfying intimate relationships—and discovering he could still get aroused and have an orgasm—Olson mustered the courage to try the online dating app Hinge. He didn’t want to be one of those stories he’d heard of prostate cancer patients just giving up on sex. Plus, he craved companionship.
On a trip to San Diego last fall to receive his immunotherapy infusion as part of his ongoing clinical trial, Olson went on what turned out to be a six-hour date with a 42-year-old schoolteacher named Ashley. After he told her about his cancer, he assumed he would be friend-zoned. But she looked past his prognosis and the thousand miles between them, and they fell in love.
Sex turned out to be sweeter and more tender, which he says is a big contrast to his former focus of concentrating on orgasms when he had more testosterone. “I love lying in bed cuddling for hours,” he says. “I never would have done that before.”
Still, it’s hard to ignore his illness. “We’ve cried together. We worry together,” he says. “I get that it’s a rollercoaster for her. She’s investing a ton into me.”
Sometimes she asks him how he sees the future. “Are you looking at five years? Or 10?”
That’s when his eyes well up, and his voice cracks. “Sometimes I let myself think, ‘Maybe it ends up being a baller 20 years,’” he says.
Then Covid-19 hit.
Olson suddenly had to deal with a new threat to his survival.
He quickly went into isolation at his house in Oceanside on the Oregon coast. Too terrified to fly to San Diego for his trial, he was able to have the oral medication— a tyrosine kinase inhibitor called cabozantinib that stops cancer from growing new blood vessels—delivered to his home. A mobile phlebotomist drew blood on his driveway in the fresh air.
But the immunotherapy atezolizumab had to be administered intravenously at an infusion center. McKay negotiated for it to be given at Oregon Health & Science University, but when Olson learned that would require him to become a patient there and lose McKay as his doctor, he decided to stop taking it temporarily.
That decision led to a sequence of events that Olson is convinced simultaneously saved and imperiled his life. First the good news: Washington Post reporter Laurie McGinley, whom he’d met on a panel during the 2018 Biden Cancer Summit, reached out to include him in a story about the challenges cancer patients faced during Covid-19.
When the San Diego ABC news affiliate picked up the story, Olson was contacted by a Good Samaritan—Shon Boney, 52, a cofounder and former CEO of the grocery store chain Sprouts Farmers Market, who was in a clinical trial at Moores Cancer Center after being diagnosed with the brain cancer glioblastoma last December. “His story hit close to home for me. I couldn’t imagine having to miss treatment,” he said. Boney, who’s a pilot, used to volunteer to fly young cancer patients and wounded warriors to medical centers several years ago and now offered to make his plane available to Olson. The gesture touched Olson deeply and made it possible for him to resume his trial.
But he missed two cycles of treatment. That was six weeks worth of medication, and he will never know if that’s what coaxed his cancer cells out of hiding. Before that, Olson’s trial regimen had stopped any new growth for 15 months.
During a trip to Montana to introduce Ashley and her 10-year-old daughter to his parents in July, Olson was practicing yoga headstands on the shores of Flathead Lake near Glacier National Park when he felt a sharp pain in his back.
The culprit was the old lesion in his T5 vertebra that had been radiated three years earlier. Soft tumor tissue had somehow broken out of the hard bone that Olson’s orthopedic surgeon likened to a charcoal briquette. The cancer’s tentacles had spread quickly and wrapped about his spinal cord. An eight-hour surgery to remove the tumor and replace the vertebra with a metal rod saved Olson from being paralyzed.
Then his surgeon did what had now become a familiar ritual: sent off the samples for sequencing.
Olson spent the next six weeks recuperating at Ashley’s house. In the evening, they grilled salmon, played backgammon under an old apricot tree in her backyard, and listened to the Yacht Rock or Rolling Stones stations on Spotify. She changed his dressings, made him acai smoothie bowls, and accompanied him on short walks around her neighborhood to build up his stamina.
The Kind Man with the Plane would fly him back to Oregon, and he and McKay would receive his genomics results. Olson feared they would contain the same disappointing collection of letters: There would be the mutations in the PI3K pathway, which had been drugged twice.
The MAP3KI variant would surely show up again, too. Olson knew there were medications to target that one, but they’re extremely toxic and don’t work well.
Then there was the CTNNB1 mutation, which his reports showed was an especially important one for his cancer. But no pharmaceutical companies had developed any drugs for it yet.
“Don’t worry,” he told Ashley. “We’ll figure something out.”
He and McKay are weighing his options.
Maybe Olson will stay on his trial and undergo extra monitoring in case other new lesions show up.
Maybe they could try focusing on the PI3K pathway again, now that a couple years had passed. Perhaps it would respond again. McKay said there was an off-label targeted drug that had been approved for breast cancer patients.
Or maybe there was a new actionable mutation or new drug combo they’d never heard about. Maybe it already had worked on someone who had the same DNA mutations and cancer progression.
Maybe that information would show up in a new sequencing report or a new database that McKay would have access to.
The clock is ticking. His twin has to be out there somewhere.