Australasian Science: Australia's authority on science since 1938

Blood Test Stops Cancer Return

Credit: Rob Byron/adobe

Credit: Rob Byron/adobe

By Jeanne Tie, Belinda Lee & Peter Gibbs

The ability to detect cancer DNA in a patient’s blood could enable doctors to predict the risk of cancer recurrence and track the success of treatments like chemotherapy in real time.

Despite advances in screening and treatment, bowel cancer continues to be a major health burden, accounting for about 700,000 deaths per year worldwide. Unfortunately, Australia has one of the highest incidences of bowel cancer in the world, and bowel cancer contributes to nearly 11% of cancer deaths in this country.

At the time of initial diagnosis approximately 80% of patients with bowel cancer have apparently curable tumours that appear to be limited to the bowel, with no evidence that it has spread to elsewhere in the body. However, following successful surgery to remove this bowel cancer, a significant minority of these patients are destined to experience cancer recurrence elsewhere in the body in the following years, reflecting the fact that the cancer must already had spread at diagnosis but it could not be detected using our current standard blood tests and scans. If these patients had been treated with chemotherapy after the surgery, about one-third of these cases would have been prevented by eradicating the microscopic residual cancer cells responsible for the cancer’s return.

In order to estimate the risk of recurrence, an examination by pathologists of the bowel cancer removed at the time of surgery is undertaken. This specifically looks for features of the cancer that are associated with a higher recurrence risk.

For example, the discovery of cancer cells in the lymph glands next to the bowel (a stage III cancer) means there is an increased probability that the cancer has already spread elsewhere (but is undetectable on standard testing). The decision to use chemotherapy is then made based on this pathology assessment.

While this is useful, it lacks precision. Ultimately, some high-risk patients will not have cancer recurrence because their cancer has been cured by surgery alone, whilst other apparently low-risk patients will suffer from recurrence.

This lack of ability to accurately determine who is most at risk of recurrence means that many patients are currently treated with 6 months of chemotherapy even though they do not need to be treated, while others that would potentially benefit from treatment do not receive the necessary chemotherapy because they appear to be at low risk.

Furthermore, for patients that are receiving chemotherapy we have no indicators as to whether or not this is being effective in the patients who do have residual microscopic disease.

Detecting Cancer DNA in the Blood

Scientists first discovered DNA circulating in human blood in 1948, but it took another 40 years to show that a fraction of the DNA circulating in a cancer patient’s blood contains cancer-specific changes or mutations that are hallmarks of cancer. Importantly, this mutated DNA is not found in the blood of healthy individuals, making it a promising biological indicator of a patient’s cancer status.

It’s now understood that when a small proportion of cancer cells rupture and die, which is constantly happening in all cancers, they release their contents, including mutated DNA fragments that float freely in the bloodstream (Fig. 1). This circulating tumour DNA (ctDNA) is then mixed in among large amounts of normal DNA fragments that are released from healthy tissues throughout the body.

The delay in the discovery of ctDNA was mainly due to the very low levels of ctDNA in the circulation, sometimes representing less than 1% of the total circulating DNA, making it very difficult to detect with older, less sensitive mutation-detection technologies. As the ctDNA levels reflect the total amount of cancer in the body, with very low levels in patients with early-stage cancers, this is particularly relevant to using ctDNA as a biomarker in early cancers.

The development of exquisitely sensitive techniques capable of detecting minute amounts of ctDNA has resulted in escalating interest in how best to use this novel blood test to improve a cancer patient’s management. For example, techniques such as BEAMing and Safe-SeqS can detect a cancer-specific DNA mutation even if it is outnumbered by normal healthy cell DNA by a factor of 10,000:1.

In 2007, Prof Bert Vogelstein’s group at The Johns Hopkins Medical Institutions first described the use of repeat ctDNA tests over time to track the cancer burden in 18 patients with bowel cancer who were being treated with surgery. They found there was a large drop in ctDNA levels after the surgical removal of all visible tumours in all patients, but in the majority of the cases the ctDNA did not disappear completely. This suggests the ongoing release of ctDNA from cancer deposits elsewhere in the body that were not detected from routine blood tests or scans.

Consistent with this, all but one of these patients with ctDNA still detected after surgery experienced cancer recurrence eventually, but none of the patients whose ctDNA became undetectable after surgery experienced recurrence. These results suggested that ctDNA could be used as quite a precise signal or marker of residual cancer cells left behind in the body after surgery, accurately predicting the risk of cancer recurrence.

This in turn could potentially be used to guide decisions about whether to recommend that an individual patient should undergo chemotherapy in an attempt to kill off the otherwise undetectable residual cancer cells that are the source of the ctDNA, and multiply to cause the recurrence that becomes apparent months or years later on CT scans or conventional blood tests.

Based on these early findings, our group at the Walter and Eliza Hall Institute of Medical Research has started several studies with Vogelstein’s group to determine how best to use this novel blood test in the day-to-day management of patients with bowel cancer.

A Proof-of-Concept Study

One of our largest studies to date is examining the value of the presence of ctDNA after potentially curative surgery for stage II bowel cancer (Fig. 2) for predicting the eventual recurrence of cancer. We know that one out of five patients with stage II bowel cancer will experience cancer recurrence despite bowel cancer surgery. As previously discussed, conventional pathology is imprecise and not very good at predicting which patients are most at risk of recurrence and which would potentially benefit from further treatment with chemotherapy. Treating every stage II cancer patient with chemotherapy makes only a small difference in outcomes, and means that a lot of patients are treated unnecessarily. A blood test that indicates the presence of microscopic cancer deposits after surgery would be extremely useful not only for guiding whether a patient needs chemotherapy after surgery, but also by measuring ctDNA levels to determine the effectiveness of chemotherapy in real time.

Our study recruited 250 patients with stage II bowel cancer from several hospitals across Australia. Blood samples for ctDNA analysis were collected 4–10 weeks after surgery and every 3 months thereafter for up to 2 years after surgery. Patients were treated routinely by their doctors without their knowledge of the ctDNA results.

The major finding from the study was the extremely high risk of recurrence for patients with detectable ctDNA after surgery compared with patients without detectable ctDNA, confirming the findings of the initial study. The risk of cancer recurrence was in excess of 80% for those with detectable ctDNA and only 8% for those without detectable ctDNA, equating to a tenfold difference in risk.

Early results also suggest that repeat ctDNA analysis during chemotherapy could be useful in assessing the effectiveness of chemotherapy. Specifically, we observed patients whose ctDNA became undetectable during or after chemotherapy who did not experience recurrence, while patients whose ctDNA level remained detectable or continued to rise despite chemotherapy eventually relapsed.

The Next Step

Despite these striking results, the biggest question remains: how much difference does a reliable indicator of microscopic cancer, or a real-time measure of chemotherapy effectiveness, actually make to saving patients’ lives (by guiding treatment toward those at high risk) or improving quality of life (by sparing low-risk patients from unnecessary chemotherapy)?

To demonstrate the ultimate clinical usefulness of the ctDNA blood test in stage II bowel cancer, we are about to conduct a randomised study where two-thirds of participants with stage II bowel cancer will be treated according to their ctDNA test results after surgery (the experimental group) and one-third of participants will be treated according to the standard recommendation made by their doctors (the control group).

In the experimental group, participants with a detectable ctDNA after surgery will receive 6 months of chemotherapy and participants without detectable ctDNA will not. In the control group, the decision about whether participants will be treated with chemotherapy is based on their doctor’s recommendation.

The rate of recurrence and the number of patients receiving chemotherapy will be compared between the two groups of participants. This will be the first study of its kind anywhere in the world where ctDNA results will be used to guide patient management.

Other Potential Clinical Uses for ctDNA

Along with use in stage II colon cancers, there are other ways that ctDNA can potentially be used in the management of patients with bowel cancer (Fig. 3).

Advanced (metastatic) cancer

Apart from determining the risk of recurrence in early-stage bowel cancer, another exciting possibility with serial ctDNA measurement in advanced cancer is the ability to watch a cancer evolve over time. Early studies indicate that repeated ctDNA testing during treatment can accurately measure the total amount of cancer in the body, meaning an assessment of whether or not a treatment is being effective can be made quite early after treatment commences rather than the current routine clinical practice of waiting until after 8–9 weeks of treatment before repeating CT imaging to measure the size of cancer deposits.

Repeat ctDNA testing over time can also help us to understand how tumours can mutate over time under the selection pressure of a targeted cancer treatment, without the need to repeatedly biopsy the cancer tissue itself. Repeated tissue biopsies can be invasive and risky. Some body organs, such as lung tissue, are not easily accessible.

Researchers have already demonstrated that ctDNA testing can be used to identify new mutations that confer resistance to therapies that a patient is receiving, and that low levels of these mutations can be found in the blood a month or two before cancer progression is detected by standard imaging techniques such as CT scans. By detecting these resistant mutations early, doctors might be able to stop a failing treatment early, reducing the use of expensive treatments that may be causing significant side-effects, and potentially switch the patient to a more effective alternative treatment.

Cancer screening

Arguably, one of the biggest impacts this promising blood test could offer to improve cancer outcomes is in the early detection of cancer. Although the current stool-based screening test for bowel cancer reduces bowel cancer incidence and death by detecting bleeding from early cancers that are otherwise unknown, the uptake in the National Bowel Cancer Screening Program is far from ideal, with the participation rate well under 50%. Further challenges are that many early cancers are not detectable with the stool test, and the great majority of patients for whom the stool test is positive turn out not to have cancer.

A simple blood test to look for ctDNA as an indicator of cancers well before they would otherwise be detected based on symptoms would be far more acceptable than the current stool-based screening test and would result in a far higher uptake in the general population. In a small study involving patients with early-stage bowel cancer (cancers that have not spread to other organs), we found that ctDNA can be detected in around 60% of cases, even in small tumours, suggesting great promise as a screening test.

Despite its many potential clinical utilities, ctDNA is not yet ready for prime time use in the clinic. These early results will need to be validated in larger studies and ultimately tested in randomised clinical trials before we can truly know its potential impact, both in terms of saving a cancer patient’s life and in avoiding unnecessary treatment and toxicity. Nonetheless, ctDNA is an invaluable research tool and has already allowed researchers to gain insights into cancer biology and evolution.

Jeanne Tie and Belinda Lee are clinician scientists and Peter Gibbs is the Laboratory Head in the Systems Biology and Personalised Medicine Division of the Walter and Eliza Hall Institute of Medical Research.