Cancer Test 'Holy Grail'? |
||
|
|
The test
The test is called DR-70, and purports to measure something called fibrinogen degradation product. This is thought to be formed in blood by the release of proteases from cancer cells. The test supposedly detects 13 different cancers from a single blood test that will cost £50, or $100.
Looking for information
The newspapers mentioned the Journal of Immunoassay. So we searched for DR-70 in this journal and found one peer reviewed paper from China [1]. No other papers were found on any search. On the Internet, we found an article about the test from the company that makes it [2]. We also downloaded and read the assay information booklet.
What is the evidence?
If you have people who do not have cancer (as far as you know), and compare them with people who do have cancer (so far as you know), then those with cancer are more likely to have higher levels of what this test is measuring. This appears to be true for a variety of cancers. The Chinese study had 277 healthy subjects and 136 cancer patients with one of 13 different cancers. A number of abstracts from the early '90s had a small number of patients.
Evaluating the evidence
Bandolier 70 described bias in diagnostic tests. The study design used to evaluate this test [1], of a group of patients already known to have the disease compared with a separate group of normal patients, was the most biased design. Bandolier 71 drew attention to evidence that 99% of 109 randomised trials from China had a positive result. Bandolier 26 also had a check list for evaluating the quality of reports of diagnostic tests: this report, if read generously, could be described as fulfilling one, and at most two, of the seven requirements.
It is interesting to note, in particular, that of the 136 cancer patients, we know
nothing about the staging of their disease, and nothing about the diagnostic criteria
used to establish the diagnosis. There is no gold standard here. Again, although 13
cancers are represented, for only three, (lung, breast and stomach) were then more than
20 patients. For nine of the cancer types represented, fewer than 10 patients were
included.
Comment
The safest conclusion is that, at best, this is early days. There is no adequate information to demonstrate the value of this test, especially in people with illnesses other than cancer, and who already take medicines. Some people who do not have cancer will have a positive result. Some people who do have cancer will have a negative result. The trouble is that no one can tell you how many, or who. The idea that this test can reliably detect 13 different cancers would be laughable if it had not made the front page of our national newspapers.
References:
- D Wu et al. Clinical performance of the AMDL DR-70 immunoassay kit for cancer detection. Journal of Immunoassay 1998 19: 63-72.
- http://www.amdl.com/IVD_article/index.html