Intervention in large bowel carcinogenesis--does screening improve prognosis?

Cancer of the large bowel is the second most common malignant tumour in the western world, with approximately 20,000 new cases registered each year in England and Wales. No improvement in the survival figures has occurred in the past 30 years, and since 1970 the number of deaths per annum in Great Britain has been rising. The prognosis of the disease is related directly to the degree of centrifugal spread of the tumour. Patients with cancer limited to the bowel wall have a corrected 5-year survival of about 90%, whereas those with tumours with lymphatic spread have a 5-year survival of about 30%. The prognosis for early colorectal tumours is therefore good, but a method of identifying such lesions is required. If early tumours bleed, the detection of occult blood in the faeces may be a valuable screening test for colorectal cancer. In one town, 8,925 asymptomatic people over the age of 40 were invited to be screened for faecal occult bleeding using Haemoccult slides. The invitation was accepted by 2,439 patients, a compliance rate of 27%. On Haemoccult screening 121 patients were positive initially, but 39 of these were found to be false-positive after further investigation. Haemoccult therefore gives a false-positive incidence of 30%. Twelve tumours of the bowel were identified, eight polyps and four cancers. Another patient who had been negative in the screening survey was identified within 2 months of the end of the experiment as having large-bowel cancer. This means that Haemoccult gives a false negative rate, in our hands, of 20%. The ability of a cancer screening procedure to identify and cure early tumours depends upon the compliance of the population, the sensitivity of the test and the availability of effective treatment. The poor compliance of our population and the disappointing sensitivity of Haemoccult mean that occult blood screening for early large bowel cancer remains theoretically possible but practically inefficient.