The Pages section of Primary Care Notebook comprises more than 30,000 concise, interlinked clinical reference articles, each focused on a specific primary care topic.
When it comes to improving the pickup rate in bowel cancer screening, how does faecal immunochemical testing (FIT), which is used in the UK bowel screening programmes, compare with the use of a blood test (looking for DNA fragments of bowel cancer)?
The bowel screening programmes in the UK are based on the use of FIT. In England, the FIT test has a positive level of 120 μg/g – i.e. a positive test is defined as there being more than 120 μg/g of haemoglobin (blood) in the stool (faecal) sample. With this current level of 120 μg/g, the proportion of patients with bowel cancer identified (sensitivity of the test) is 47.8%. Therefore, more than half of people with bowel cancer will not be identified by FIT at the current level.
If the level of FIT is reduced to say 40 μg/g, then the sensitivity of the test increases to 71.1% (and so only just under three in every ten people with bowel cancer will be “missed”). If the FIT level is 20 μg/g, then the sensitivity increases to about 80%.
The problem with reducing the thresholds is that the lower the level the lower the positive predictive value (PPV) of a positive test – i.e. the likelihood that a positive result will mean a diagnosis of bowel cancer. The lower the PPV, the higher the number of people who will have unnecessary tests and the greater the stress on investigation capacity. For example, for a positive FIT result at a level of 120 μg/g, about one in seven individuals would have bowel cancer or a high-risk precancerous lesion. In contrast, for a positive FIT at a level of 20 μg/g, one in 25 would have a cancer (24 out of 25 individuals investigated would not have a cancer; PPV, 0.04).
So how does the recent trial of a blood test looking at fragments of DNA of bowel cancer compare? The evidence suggests that the sensitivity is impressive, at about 83% (so the blood test will identify about eight of every ten individuals who have colorectal cancer in the screening population). Possibly more impressive, however, is the PPV, which is about 90% (i.e. nine out of ten people with a positive test will have bowel cancer; PPV, 0.9).
The blood test is not so good at picking up precancerous lesions as FIT, but this is because it is not looking for DNA associated with these but only DNA of cancers.
This blood test seems a potentially very important addition to the armoury in diagnosing bowel cancer.
For more information, see Primary Care Notebook.
Other highlights in this month’s email include evidence relating to stopping smoking and reduction in the risk of head and neck cancers, more evidence on the effectiveness of potassium-competitive acid blockers, data on cranberry juice in the prevention of urinary tract infections and an update of the adult polycystic kidney disease section.
Since 2001, GPnotebook has been a "go-to" clinical reference source for busy primary care professionals seeking quick answers to clinical queries
Written by practising GPs, GPnotebook comprises more than 30,000 pages of clinical reference material, with each article focused on a specific primary care topic
Information is organised into short, easy-to-read articles – quick enough to use during consultations. Hundreds of updates every month
As well as the written reference material, we have a growing library of video and podcast content, along with educational modules and quizzes to support learning and personal development