In late August 2014, Marcus Krause, a 53-year-old photographer from Atlanta, received a phone call from his oncologist. Krause had been waiting for this call for more than two weeks. He was to be informed whether his lung cancer was treatable or not.
Marcus Krause's condition had been relatively asymptomatic. Looking back, however, there had been some signs that he didn't pick up. For instance, he had been oblivious to a persistent dry cough he had developed. He had also been feeling slightly out of shape - he regularly cycled long distances but lately it would take him longer to recover. Until one day, in mid-July, when he was forced to stop abruptly during a bike sprint to the top of a hill. He felt out of breath and had to sit down for several minutes to recover. In the next few days, he developed flu-like symptoms and went to see his doctor. An X-ray revealed that his right lung was being compressed by pleural fluid. The following morning, the doctor extracted a few litres of liquid.
"They couldn't remove all of it at once because my body had grown accustomed to having that fluid in there," Krause says. The doctor installed a catheter on his ribs and instructed his wife, Amy, on how to drain the rest of the fluid at home. A couple of days later, Krause underwent a PET scan. It revealed the culprit for the pleural effusion: a 2.8cm tumour lodged on his bronchial tube. Given the location of the tumour, performing a tissue biopsy would have been complicated, so the oncologist collected tumour cells from the pleural fluid. The diagnosis: Krause had stage-four lung adenocarcinoma. In other words, lung cancer.
The question now was whether his cancer had a gene mutation that could be targeted by existing drugs or not. Targeted cancer therapies interfere with the genetic mutation that causes the tumour to grow. They outperform more traditional treatments such as chemotherapy and are generally effective if you have the right drug for the specific cancer mutation. Of the 150 genetic mutations known to cause cancer, 70 are currently treatable by targeted drug therapies.
Krause had done his homework and thought it was likely that for someone in his fifties and healthy, who had never smoked, he would have a targetable mutation. When he finally got a call from his oncologist, however, he received some bad news. There was nothing they could target.
Initially, Krause responded well to chemotherapy. "It was a new treatment that hadn't yet been approved, but it was being fast-tracked," he says. "My oncologist really wanted me to do that and made an appeal to the insurance company on my behalf, so they granted me an exception." After 12 weeks of chemotherapy, his first CAT scan showed a 70 per cent tumour reduction. His second scan in December 2015, however, showed that it had grown to five centimetres. Krause's wife cried when she heard the news. Krause, on the other hand, felt relieved. "I didn't want more chemotherapy," he says. "It just makes you so sick and I knew that it wasn't going to cure me. So the question was: do I just want to feel like hell and then die, or do I want to maybe feel OK and then die?"
At this point, Krause's doctor referred him to an oncologist at the Sarah Cannon Research Institute in Nashville, in the hope that he could at least enrol him in upcoming clinical trials. The Nashville oncologist, however, wasn't convinced about Krause's biopsy results and asked him if he was willing to undergo another test. It wouldn't be a tissue biopsy, but a blood test instead. The test was new but early studies had shown promising results. Unlike a tissue biopsy, it wasn't invasive. All they needed was two vials of blood. Krause had nothing to lose. He agreed.
In the meantime, he was put on another course of chemotherapy. He was busy at work, his eldest daughter's wedding was coming up and he needed something to keep him going. "The wedding was emotional," Krause says. "Everybody knew things were not looking good at that point." The new chemotherapy soon also proved ineffective. By then, Krause had forgotten about the blood biopsy test. "When I called my oncologist to ask 'What now?', he said, 'You remember that blood test we did?'" The test had revealed a mutation, an epidermal growth factor receptor (EGFR), that the first biopsy had missed. Krause had reason to hope again.
One day in 2013, Helmy Eltoukhy, a 34-year-old electrical engineer from California, flew to Spain to meet a Swiss researcher named Maurice Stroun. Eltoukhy had just started his second company, Guardant Health, with former PhD colleague AmirAli Talasaz, but he didn't have a product yet. Instead, he had a vision - to invent a new test that would replace tissue biopsies with a blood test. And Stroun, one of the scientists who had discovered the existence of cancer DNA in blood, could help him.
Eltoukhy had studied at Stanford University, where he was involved in projects that used semiconductors to make cheaper and faster DNA sequencers. His group, headed by Iranian molecular biologist Mostafa Ronaghi, had received $10 million (£7.7m) in grants. With that money, Eltoukhy and Ronaghi spun out a company, Avantome, in 2007. A year later, they sold Avantome to Illumina, the world's leading manufacturer of sequencing machines. Eltoukhy and Ronaghi joined the company, where they supervised half of the projects at its R&D department.
In 2011, Eltoukhy felt ill. The list of symptoms was unusual: brain fog, arthritis, stomach pains and exhaustion. He went to 12 specialists, undergoing multiple blood tests, a colonoscopy, an endoscopy and a CT scan. No one could tell him what was wrong with him. "I was freaking out," he admits.
One evening, his wife organised a make-your-own-pizza party at home for their friends. Eltoukhy binged on pizza and bagels and, later that night, a rash broke out, covering his entire body. After seeing him, his doctor suggested he might want to try cutting gluten out of his diet. Four weeks later, his health was restored. The appointments with various physicians during those six months made Eltoukhy reflect that they had poor technological tools to help in diagnosing and treating disease. "I always considered physicians to be the heroes on the front line," Eltoukhy says. "They are on call, they are helping people at their time of most desperate need and yet they don't have the tools to ask and answer the right questions."
Eltoukhy decided to quit his job and do something about it. "It was all about reducing the cost of sequencing," he says. "Can we get to the $500 genome? The $250 genome? I felt we were killing ourselves with advanced physics while there was lots that we could be doing closer to patient care. I felt that if I wasn't a part of this revolution I would regret it." He decided to focus on the disease that had taken his grandmother at 40, and is still one of the world's biggest killers. "Cancer is a disease of the genome," adds Eltoukhy. "I figured that's the one area where our knowledge of advanced genomic sequencing would make a big difference."
From his perspective as an engineer, how oncologists diagnose and treat cancer constitutes what Eltoukhy considers a faulty feedback loop. "When you think about diseases that we have been able to combat effectively, whether infectious diseases, HIV or even blood cancers such as leukaemia and lymphoma, which have fairly high survival rates now, it's because we have access to information about the disease through a simple blood test," Eltoukhy says. "We can access and sample the disease continuously. Unfortunately, in solid tumour cancers, it takes three to nine months to figure out if a drug works. The feedback loop is enormously slow."
Furthermore, this slow feedback loop doesn't even track the underlying genomics of the disease. A cancer genome evolves all the time - the early mutation that drives the initial stages of the cancer will not be the same a few years down the line. Its genome can change, adapt to treatments and multiply into various mutations that evolve independently from the original one. "Cancer is the one disease where you will do an initial tissue biopsy and use that information for the rest of that patient's life," Eltoukhy says. "By then the cancer has grown, acquiring different mutations, and the physician may have no idea. It could be five or ten years, and we're still acting on the original diagnostic. Unless you keep testing, you're not going to be able to effectively correct the treatment and change the drugs every time it's required. We have to manage the disease dynamically and adaptively."
He soon found out that a blood test could provide one possible way to measure cancer mutations without the need for an invasive tissue biopsy. It would allow doctors to test patients more regularly to track the cancer's evolution. This would avoid the need for several painful and invasive tissue biopsies that were detrimental to the health of the patient. It could also be available as part of an annual checkup, something that would work equally for early diagnosis and the sickest patients. When he searched for articles on cancer DNA in the circulatory system, all the references pointed towards one name: Maurice Stroun.
Stroun wasn't a cancer expert. He was a plant physiologist at the University of Geneva. In the 60s, he had studied tumours caused by bacteria in plants and found genetic material of the bacteria in the sap of tomato plants. This led him to study animal species and, in 1972, he found bacterial DNA in the circulatory system of frogs. This showed Stroun that it was possible to detect genetic material from a foreign body such as a bacterial tumour in an organism's circulatory system. When he came across a study that showed that cancer patients possessed higher quantities of DNA in their blood serum than non-patients, Stroun wondered whether that extra DNA was from the cancer tumour itself. In 1994, his group became one of the first to successfully detect DNA from tumours, so called cell-free DNA, in cancer patients' cardiovascular systems.
"The theory was that, as a cancerous tumour grows and its cells multiply, it begins outgrowing its blood supply," Eltoukhy says. "Even in early-stage cancers, its cellular growth rate is only narrowly higher than its death rate. As cells die, they shed their contents into the blood stream, including its DNA."
When Eltoukhy met Stroun, they immediately hit it off. Guardant Health subsequently acquired several of Stroun's patents and he became the company's founding adviser. In 2014, Guardant launched its first blood biopsy test. "Stroun's a guy who didn't get a lot of credit and respect from the cancer community, because of his background as a plant physiologist," Eltoukhy says. "Of course, a lot of the sequencing technology had to be developed to make comprehensive liquid biopsies happen. I think people are now looking back and realising he was right about a lot of things."
It's July 15, 2016, and WIRED is attending a meeting at Guardant Health's offices in Redwood, California. Some challenging case studies are discussed. Richard Lanman, Guardant's chief medical officer, leads the group. "Every week, new employees join the company," Lanman says. "Reviewing cases is a way to introduce them to our technology."
Lanman is in his fifties and has been at Guardant since September 2014. One of the first studies he conducted was an analysis of the costs of biopsies which was published in the Clinical Lung Cancer journal. The study reported that a lung biopsy cost, on average, $14,634. This included the cost of complications from the surgery, experienced by one in five patients who undertake the procedure. Guardant Health's test costs $5,800.
"Today's physicians want more tissue, so the needles are getting bigger, which only increases the complication rate," Lanman says. "If you go to a top hospital you should have no problem getting a comprehensive sequencing of your tumour. But what about patients who don't have such access?" In the US, Guardant Health bills health insurers directly for the test. "We're democratising access to biopsies."
A few weeks before WIRED's visit, at the annual American Society of Clinical Oncology meeting, Helmy Eltoukhy had announced the results of a study involving 15,000 cancer patients. The study demonstrated that Guardant's blood-test results agreed 99 per cent of the time with the results of a standard invasive tissue biopsy. This concordance level dropped to 67 per cent when a Guardant test was compared to tissue biopsies taken more than six months earlier. "That's what you would expect," Eltoukhy tells WIRED. "It tells you that the original mutation has evolved and the original biopsy result no longer applies."
This secondary wave of mutations is what Lanman calls "the landscape of resistance". "These mutations are why treatments almost always fail," he says. The first genetic mutation that leads to cancer might be caused by a variety of factors: environmental, chemical, gamma radiations from space, viruses or inherited defects that impair our ability to repair DNA. "When enough of those mutations happen in the right way it causes cells to go out of control," he says. This initial mutation will continue to grow unimpeded until it's resisted by medical treatment. "Until then, cancer cells had no reason to mutate," Lanman continues. "Treatment forces the cancer to evolve, to get around it."
It's this landscape of resistance that Guardant is starting to map. The more clinicians know about the type of mutations that arise after the first mutation, the better these mutations can be targeted throughout a patient's lifetime. "Because our blood test is used most often when the first line of treatment has failed, we have the largest database of resistance mutations," Lanman says.
His hope is not for a cure, but for management of the disease much as we manage HIV. "People have talked about cures since 1980," Lanman explains. "I don't think that will happen any time soon. I do, however, think we're on the verge of managing cancer."
After his introduction at the meeting, Lanman presents a few case studies. The first was a patient - male, 55 years old, based in Tel Aviv - with metastatic lung cancer. They found an EGFR L858R mutation for which there is a match therapy, a drug called erlotinib. "He did well for a while, but then the cancer evolved," Lanman says. "When they biopsied the patient again, 13 months later, they found eight new mutations." None had a treatment at the time.
In July 2014, the patient took a blood test with Guardant. The test found another mutation, T790M, in the EGFR gene. "It was driving all the resistance to the treatments," Lanman explains. "We tried to get into a clinical trial for a new drug for that mutation, but the company didn't enrol him." Guardant had just launched and the pharmaceutical company was unwilling to take a risk on a new diagnostic tool. A few months later, the patient's tumour metastasised, invading the liver. In September 2014, a standard liver tissue biopsy detected the EGFR-T790M mutation, confirming Guardant's result. They put him on a clinical trial and, by January 2015, the tumour had shrunk by 40 per cent. "He did well until they did another lung tissue biopsy. They stuck a needle in him, caused a lung collapse, and he died." The biopsy - not the cancer - killed the patient.
Another case study presented by Lanman was a 60-year-old male patient with metastatic colorectal cancer. The patient had received treatment, but it was stopped when he developed complications. "He was frail and ready for palliative care," Lanman says. The tissue biopsy, however, hadn't detected an ERBB2 gene amplification, an alteration in the number of copies of the gene. This is one of four types of cancer mutations. The basic one is just a letter change in the four-letter genomic code. The second type is a fusion of two genes, and a third happens when a letter is inserted or deleted from the code. The fourth type is an alteration on the number of copies for each gene, of which the ERBB2 amplification is one example. "We all have two copies of every gene in our genome," Lanman says. "In some cancers, you can have eight, 15 or 100 copies. These can be hard to detect."
Guardant's test can detect the four types of mutations. The patient was put on trastuzumab, an extremely efficient drug. "The tumour dissolved," Lanman says. "The results can be dramatic even when you are on death's door."
A final case study: a 58-year-old female with lung cancer and bone metastasis. "She received chemo and three biopsies," Lanman explains. "But they couldn't get enough tissue to sequence." Her oncologist ordered a blood test; the results showed the presence of two molecules with a EML4-ALK fusion mutation in ten millilitres of blood. "It was a low amount," Lanman says. "But if it's there, it's a detection."
Guardant's blood biopsies have a specificity of 99.999999 per cent - which means that the test reports virtually no false positives - thanks to a digital sequencing method it developed. "We barcode the two complementary strands of DNA and ensure they match. Most sequencing methods don't include this step, and can get false positives. There's nothing like this in diagnostics," Lanman says.
The patient was treated with a drug called crizotinib and recovered. "She went from no options to a great response to treatment because of a blood test."
After his Guardant test in 2015, Marcus Krause begun a treatment with erlotinib. Before, he had to take narcotics to alleviate the pain. Within two days of starting on the drug, he no longer felt pain. After eight weeks, he took a CAT scan. His previous scan had shown a five-centimetre tumour, with nodules invading his chest cavity and a cancerous lymph node on his liver. The new scan showed that the lymph node had cleared, the nodules were gone and the tumour had shrunk to one centimetre.
In May 2016, at the invitation of Helmy Eltoukhy, Krause visited Guardant's offices in Redwood to meet the team and share his story. That day Guardant was launching a new initiative called Project Lunar in collaboration with the University of California, San Francisco, South Korea's Samsung Medical Centre, the University of Pennsylvania and the University of Colorado. The study involves, at an early stage, hundreds of pre-symptomatic, high-risk individuals for breast, ovarian, colorectal, pancreatic and lung cancer. "These are patients who, like Angelina Jolie, have a BRCA mutation for breast cancer," Eltoukhy says. "Most of these women have to make a choice about removing their breasts and ovaries. What the study will do with those having surgery is to take a blood sample and compare it with the tissue that has been removed, where sometimes you find early stage-cancer lesions."
With Lunar, Guardant is focusing not on the landscape of resistance, but on the original mutation. If successful, the study will be a step in the direction of Eltoukhy's original vision: a tool for early detection of cancer. "They will be able to use our test for active surveillance rather than make a big decision to surgically remove their organs," Eltoukhy says.
That afternoon, Krause took part in a panel discussion. He recounted his story to an audience of Guardant scientists and lab technicians: how his mutation had been missed with a standard tissue biopsy; the bad shape he was in when the chemotherapy stopped having an effect; and how the blood test had saved him. He knows that erlotinib won't last forever. At some point, the cancer will mutate and become resistant to the drug. There's a 60 per cent chance that the new mutation will be T790M, in which case the prescribed treatment will be a drug called osimertinib. He can live like this, he says. He feels like he's not even sick.
"My story made such a big impression," Krause remembers. "The scientists spend most of their time in the lab and to hear such an incredible turnaround story from a person standing in front of them was a big deal." At the end, he told the audience: "When things get rough again, I'm just going to remember standing here looking at each of you. That's going to inspire me to keep going."
This article was originally published by WIRED UK
