Need a liquid biopsy test?
Recent news reports suggest that you do… but do you really?
These reports are based on a paper published in Science on 18th January 2018 (including Australian co-authors) proposing that there may well be a benefit from liquid biopsies (“Detection and localisation of surgically resectable cancers with a multi-analyte blood test”). Cohen and his colleagues state that the 8 cancer types they studied accounted for 60% of cancer deaths in the USA in 2017 and that earlier detection could conceivably reduce deaths from these diseases. The media is now postulating that executives might have an annual blood test for circulating tumour DNA (liquid biopsies) as part of their annual physical. The authors used a combination of ctDNA and proteins in the blood to test 1005 patients with early cancer and no obvious secondaries.
The field of genomic medicine with companion targeted therapy is taking off with “genomics” becoming a buzz word and usually referring to testing for mutations from the primary or secondary cancer. This testing usually has involved a tissue biopsy. In certain cancers it has made an amazing difference in treatment (for example NSCLC and EGFR mutations – see my blog on EGFR mutations in lung cancer), both initially and at relapse. In colon cancer the presence of one mutation (the KRAS mutation) signals that one particular type of chemotherapy/antibody treatment will not be useful, thus saving the time and expense of prescribing it and monitoring patient response. More recently, genomic testing has become available on circulating tumour cells (cells within the blood stream). This is called ctDNA and is helpful when tissue is not available or when multiple samples need to be taken but multiple tissue biopsies are inappropriate (e.g., due to access issues or risks related to repeated biopsies). ctDNA testing is now available in Australia, as well as in the USA, for EGFR mutation testing in lung cancer if a specimen cannot be obtained by a normal biopsy.
Are there any problems with ctDNA testing?
ctDNA secretion and tumour size
- Cohen et al. report that not all cancers secrete cancer DNA in to the bloodstream and even those that do only secrete when over a certain size. Hence claims that liquid biopsies will identify early cancers may be exaggerated for some cancers. Editorial Science Jan 2018
- The patients in the study were already known to have cancer, therefore they would have had enough symptoms to draw themselves to the attention of their doctors and to have had their diagnoses made. They were early stage patients generally with no evidence of metastatic disease. It is not known how the test will perform in patients with earlier non-symptomatic disease. This raises the issue of false positives discussed below.
False positives in cancer testing
- The liquid biopsy or CanSEEk test study looked at ctDNA and proteins in 812 individuals with no known cancer and 1005 patients with KNOWN early stage cancer of 8 specific types (ovary, liver, stomach, pancreas, oesophagus, colorectal, breast and lung). There are reasonable screening techniques for the last three, but not the first five. CanSEEK tests for a number of protein markers, some of which are commonly done and others which are not. The authors state that the test has a specificity of 99% which means that of the 812 people without the disease, it correctly excluded 805 of them. While this sounds wonderful, it needs to be considered more carefully.
The results for the 7 patients (roughly 1%) who did not have cancer are known as false positives. That means they don’t have the cancer but the test suggests that they do (see diagram). The issue with false positives is that when large scale screening is undertaken the number of people who screen positive remains at 1%. The way this is represented is shown below. The group who do not have the disease but test positive will be subjected to many more tests (expensive and possibly uncomfortable) to make sure they do not have cancer.
- Consider what might happen if large numbers of patients who have no symptoms go to their doctors for liquid biopsy tests to see whether they have cancer. Let’s take colorectal cancer (CRC) as an example. In 2017 it was anticipated that approximately 16,000 new cases of CRC would be discovered in Australia. Approximately 50% of patients with CRC have no symptoms and so 8,000 patients would be symptom free. Presently, the standard techniques of faecal occult blood screening offered by the government and regular colonoscopies should detect most of these. About 300,000 Australians per year have the government test (a participation rate of 37%).
- Assume that double this number would sign up for a “simple blood test” instead of undergoing faecal occult blood testing. This would mean around 600,000 Australians per year might ask for this blood test. The specificity of the test is 99% which means that the false positive rate (100% minus specificity) is 1%. Using the liquid biopsy test, 6,000 screened patients would have a false positive test. (False positive rate of the test is 1% of 600,000). These individuals would all need further testing to exclude cancer.
- One follow up study is suggesting that a positive test be repeated and only after 2 positive tests would further investigations be done. Even so, a large number of people without cancer are likely to experience anxiety while undergoing a series of blood tests and investigative procedures. As colorectal cancer is just one of the eight cancers identified for liquid biopsy, the situation is compounded in screening large numbers for all of the eight cancers.
There are some real concerns around the issues of test specificity and false positives. These caused much anguish in the early days of the AIDs epidemic when a very specific test for HIV/AIDS was rolled out in to a community that did not have a high number of patients with the disease. Many individuals who received false positive results experienced high levels of unnecessary anxiety.
Detecting a cancer relapse with ctDNA
In patients who have a known cancer this test and others like it using a combination of ctDNA and protein testing may be useful for monitoring and detecting early recurrence of the tumour.
What is the best use of CanSEEK?
- One problem with tissue mutation testing on a biopsy specimen is that the biopsy specimen may be small and after all of the conventional testing, very few cancer cells may remain for extraction of DNA for mutation testing. In the cases of insufficient tissue specimen, if DNA is released into the circulation, then it may be detected by liquid biopsies.
- It is suggested that the technique could be useful for diagnosis of cancers that do not have an established screening method. In over 90% of patients with known ovarian and liver cancers, CanSEEK identified the cancers and in over 83% of patients, a primary site was adequately established. However, it is unknown how successful the test is in identifying these cancers in an apparently well population, i.e., those who present with no or vague symptomatology.
This test (CanSEEK) may well offer a significant advance, but only if it is used with knowledge and care and not as an indiscriminate screen for every normal person to check whether they have cancer. The test has been mentioned in many newspapers recently, but few, if any, report the possible danger of false positives. General practitioners should be alert to the fact that they may see an increase in patients asking about the test due to the media coverage. An algorithm to decide which patients would benefit and a means of educating patients before they have the test would be of benefit.
Please note that the information contained in this article is for educational purposes only. It is important that you seek medical advice from a medical practitioner or specialist (e.g., Oncologist) on the most appropriate treatment for your cancer.
If you are unsure of the tests available for your cancer or what your treatment options are, I am available to discuss these with you.
Contact Dr Hawson for an appointment