快猫短视频

CASEY LIM (not her real name) was 11 weeks pregnant when she gave a blood sample to test for Down鈥檚 syndrome. It revealed multiple abnormalities in fragments of DNA circulating in her blood. Further analysis showed that the mutant DNA wasn鈥檛 from the fetus but from Lim鈥檚 own white blood cells. A form of blood cancer called follicular lymphoma, which she had battled two-and-a-half years earlier, had stealthily returned.

Lim was one of the lucky ones. All too often, cancers are caught only after they have spread and mutated into subtly different forms. This ability to evolve and resist our best treatments is what makes cancer such a formidable foe. One way to root out its weaknesses is to repeatedly remove chunks of tumour for analysis, but that is invasive and often risky.

So how do we keep an eye on a cancer鈥檚 ominous transformations, figure out the best way to treat it and check that it is not stealthily growing back? The answer could be written in blood. A test based on tiny bits of DNA released into the circulation by cancer cells might even give doctors a way to catch the disease earlier.

鈥淭he best time to treat cancers is when there are very few cancer cells, when there are fewer chances for them to develop resistance,鈥� says , director of the Ludwig Center for Cancer Genetics and Therapeutics at Johns Hopkins University in Baltimore, Maryland.

Vogelstein鈥檚 lab is one of several dedicated to developing a blood test to spot cancer early and follow its every move. It鈥檚 a laudable ambition. To make these 鈥渓iquid biopsies鈥� reliable enough to be useful, however, they will have to figure out how to track down and decode the faint, potentially misleading traces tumours leave in the blood.

All cells, including cancer cells, release DNA into the blood when they die. Researchers have known for some time that it is possible to detect this free-floating DNA in the circulation, and that people with cancer have more of it than healthy people. In the earliest stages of the disease, the number of fragments deriving from cancer cells 鈥� known as circulating tumour DNA, or ctDNA 鈥� is vanishingly small. But as the tumour grows, so does the quantity of DNA fragments.

In the late 1990s, , now at the Chinese University of Hong Kong, was struck by the opportunities this offered. His main interest was in developing non-invasive ways to detect fetal abnormalities. Eventually, he came up with a way to detect fetal DNA in the blood plasma of the mother, and within 10 years he鈥檇 devised a blood test for Down鈥檚 syndrome. He also set his sights on cancer.

Tumour DNA is rarer than fetal DNA, but as sequencing technology has improved, so has our ability to catch it and read what it can tell us about the cancer. The simplest strategy is to search for genetic mutations you already know about from a biopsy of the tumour.

In 2007, Vogelstein鈥檚 group used this approach to monitor tumours in 18 people with bowel cancer. Their technique took advantage of the fact that bits of single-stranded DNA stick to other bits with the corresponding sequence of nucleotides. They made fragments of DNA that mirrored the mutation they were trying to detect, which were tagged with a fluorescent molecule.聽They also figured out a way to copy individual pieces of DNA in a blood sample onto beads, where the DNA could be amplified to produce many individual copies. Then the fluorescent fragments were added to the beads, and any glowing beads counted to give an聽estimate of the concentration of ctDNA in the blood.

After these patients had their tumours removed surgically, a blood test was still able to detect in most of them, even though quantities had fallen dramatically 鈥� and all of those people eventually saw their cancer return. In the handful of patients with no detectable tumour DNA in their blood, the cancer has not come back. This suggests that the quantity of ctDNA in blood reflects tumour size, and that liquid biopsies could be a useful monitoring tool.

In England, 46 per cent of cancers are diagnosed at an advanced stage

Since then, Vogelstein has shown that almost every type of cancer sheds fragments of DNA into the blood. In 2014, his group announced the detection of ctDNA in the blood of more than 75 per cent of a group of 640 people with a range of cancers, such as advanced pancreatic, ovarian, colorectal and breast cancer. Brain, renal and prostate cancer seemed to shed fewer DNA fragments 鈥� although they could still be detected in half of those whose disease was at an advanced stage (, vol 6, p 224).

Shards of DNA can also be used to predict whether those with cancer will respond to drugs. Last year, for instance, the European Medicines Agency approved a blood test that indicates whether people with non-small-cell lung cancer will respond to gefitinib, a drug which works against cells carrying a particular mutation. This is a big deal because taking a lung biopsy is difficult, meaning doctors often don鈥檛 get enough tissue to understand the tumour. 鈥淭hat鈥檚 a perfect situation for a liquid biopsy,鈥� says Vogelstein.

You also want to look at how the cancer evolves to evade treatment. One of cancer鈥檚 grim ironies is that the better the drug you throw at it, the more you prime it to resist. 鈥淪ome of the targeted drugs against cancer are so dramatically effective that they select for major changes in the cancer biology,鈥� says , director of the Massachusetts General Hospital Cancer Center in Boston. 鈥淭his is a great rationale for monitoring cancers through the blood continuously.鈥�

Vanishingly rare

At the Institute for Cancer Research in London, Delila Gasi-Tandefelt and her colleagues have done just this, taking sequential聽blood samples from men with advanced prostate cancer. Their aim was to look for changes in a known genetic sequence that predict how likely it is that the patient will respond to the drug abiraterone. Sure enough, the researchers , such as CT scans. 鈥淚nstead of giving them abiraterone, which they probably won鈥檛 benefit from, we can give them an alternative treatment,鈥� says Gasi-Tandefelt.

Impressive as these early applications of liquid biopsies are, they are merely the low-hanging fruit. 鈥淎 major interest all along has been to develop a diagnostic test that could be applied to asymptomatic people,鈥� says Vogelstein. Considering that even in wealthy countries nearly half of cancers are diagnosed at a late stage, a blood test that could detect the disease early could save millions of lives.

The devil is in deciding what to look for. If you already have a sample of the cancer tissue, it鈥檚 relatively easy to probe the blood for its signature mutations. If not, you need to look for all kinds of mutations. What鈥檚 more, ctDNA is exceedingly rare in the early stages of cancer. 鈥淭he problem with blood is there鈥檚 a lot of stuff in it, and the material from cancer is relatively small,鈥� says Haber.

Another approach is to sequence all of the circulating DNA in the blood and then look for differences in the ratios of fragments from various regions of chromosomes, the thread-like structures in cells that contain most of your DNA. This might flag up whether some areas have become duplicated, deleted or rearranged 鈥� all classic hallmarks of cancer.

In 2012, Vogelstein鈥檚 group sequenced DNA from the blood of 10 people with late-stage breast or colorectal cancer and 10 healthy people, searching for these structural alterations. They detected abnormalities in all of the cancer patients, but none of the healthy people. However, it is not certain that the technique would work in cancer鈥檚 early stages, when there are fewer bits of rogue DNA to analyse. Nor do we know if it would detect every significant alteration, or distinguish between pernicious and harmless changes.

The caveats don鈥檛 end there. Even a test that is sensitive and broad enough to detect very early cancers runs the risk of generating false positives (see 鈥�When is a cancer not a cancer?鈥�) unless you can confirm the result with some sort of imaging technique.

And assuming you can detect the genetic signature of a cancer, how would you know which part of the body is affected? Some mutations provide a clue: a mutant KRAS gene, for instance, is usually associated with lung, colon or pancreatic cancer. Other mutations, such as in the p53 gene, could be associated with just about any form of the disease.

In 2014, just 12.8 per cent of the US National Cancer Institute鈥檚 research budget was spent on detection and diagnosis

The good news is that ctDNA isn鈥檛 the only tumour-derived material in the blood. There are also circulating tumour cells. They are harder to spot than ctDNA as there is typically just one CTC for every billion blood cells, and they may not get into the bloodstream until later in the cancer鈥檚 development. It is possible to catch them, though, either by separating them by size from smaller white blood cells or fishing them out using magnetically tagged antibodies that latch onto specific cell-surface markers. And CTCs can reveal far more than ctDNA.

With CTCs you can scrutinise not only the tumour鈥檚 DNA but also its RNA 鈥� the chemical go-between that helps translate genetic code into proteins. This extra information ups the chances of pinpointing the site of the cancer and offer clues about the genetic reshuffling that might be driving it, such as translocations 鈥� where a chunk of DNA has jumped to a different location in the chromosome, potentially altering the expression of genes nearby. Such insights could be vital in diagnosing prostate cancer, for example, which is more often characterised by translocations than a particular mutation.

Another option is to focus on platelets, the sticky discs in blood which promote wound healing by forming clots. As they circulate, platelets accumulate fragments of RNA from the tissues they come into contact with, including cancer cells. By sequencing this RNA, you can learn about tumours in their earliest incarnations, long before CTCs can be detected. Indeed, Thomas Wurdinger at the VU University Medical Center in Amsterdam, and his colleagues recently . They were also able identify the location of the primary tumour in 71 per cent of cases.

In the US, only 15 per cent of lung cancers are caught before the disease has spread elsewhere

Then there are exosomes, tiny secreted packages containing DNA, RNA and proteins. Last year, researchers at the University of Texas MD Cancer Center in Houston were able to .

So ctDNA is not the only game in town. It remains a strong contender, though, and methods for reading abnormal DNA fragments are improving all the time. Lo鈥檚 group, for instance, has developed a novel technique that looks at methylation 鈥� where chemical tags known as methyl groups stick to DNA, often altering gene expression. This allows them to trace ctDNA back to its tissue of origin by comparing its methylation pattern against a database of tissue types.

Casey Lim鈥檚 cancer was spotted because she gave a blood sample to Lo, who was testing this new approach. It traced the abnormal DNA to her immune cells. 鈥淐hemotherapy had to be commenced immediately,鈥� says Lo. 鈥淕iven the early stage of pregnancy, the patient opted for a termination, received the chemotherapy and is currently well.鈥�

Lo and his colleagues also recently discovered that DNA fragments from tumours tend to be significantly shorter or longer, on average, than those from healthy cells. They showed that is enough to distinguish people with liver cancer from healthy people, even at a very early stage of the disease.

As the sensitivity of such tests improves, they could start to make a real difference. Vogelstein points out that on average it takes 20 to 30 years for a cancer to progress from initial mutations to a full-blown metastatic disease. That is a big window in which to intervene, one that we are failing to exploit.

鈥淚n the US this year there will be about 55,000 people who die from colorectal cancer, and every one of them will only have died because their cancers were not detected in the first 27 years or so,鈥� says Vogelstein. So even if they can鈥檛 reliably detect cancer in its earliest stages, blood-based biopsies could still give us a valuable head start.

This article appeared in print under the headline 鈥淲ritten in blood鈥�