Premalignant conditions of the gastrointestinal tract: a basis for selective screening programs

Extracolonic manifestations of familial adenomatous polyposis: desmoid tumours, and upper gastrointestinal adenomas and carcinomas

It is well known that patients with familial adenomatous polyposis (FAP) are at considerable risk of developing extracolonic manifestations of the disease. Particularly, desmoid tumours of the abdominal cavity, and duodenal adenomas and carcinomas are the most serious ones. It is estimated that some 10% of the FAP patients will develop desmoids, whereas 50-90% of the FAP patients will get duodenal adenomas predominantly concentrated on or around the major papilla. Desmoid tumours and duodenal carcinomas are major causes of death in those patients in whom a prophylactic (procto)colectomy has been performed. Desmoids are histologically benign tumours, composed of mature fibroblasts. They usually grow slowly but they can become quite large and may compress or infiltrate surrounding viscera, which might cause significant morbidity as well as mortality. Successful treatment of these tumours is extremely difficult as surgical therapy often requires the removal of considerable lengths of small bowel. Moreover, surgical therapy may lead to uncontrollable bleeding and is seldom radical. Chemotherapy with cytoxic agents seems promising but so far the data are too few for firm conclusions to be drawn. The same holds true for drug regimens which interfere with the metabolic and hormonal metabolism of the tumour. Although various lines of evidence suggest that the adenoma-carcinoma sequence, which is generally accepted for colorectal adenomas, also applies for the duodenal adenomas in FAP patients, it is not clear whether we should screen these patients for upper gastrointestinal adenomas or not. As these polyps are usually small, sessile, multiple and difficult to remove, the benefit of endoscopic surveillance would be the early detection of cancer rather than eradication of the polyps. In addition, evidence that screening and early treatment leads to improvement of the prognosis is not available. Although the role of (procto)colectomy in the treatment of large-bowel polyps is well established in FAP patients, the treatment of their duodenal counterparts is still open for debate. The risk of the development of periampullary cancer is not high enough to warrant an aggressive prophylactic surgical approach, i.e. a Whipple's procedure, immediately after the discovery of duodenal adenomas. The considerable morbidity and mortality rates of this procedure must be weighted against a putative benefit of screening.

How prevalent is cancer family syndrome?

Based on an established but pragmatic definition of cancer family syndrome as the presence of three or more relatives affected by colorectal cancer in a first degree kinship, the contribution of this syndrome to the total cancer burden in Northern Ireland has been studied by investigating all non-polyposis probands under 55 years old at histological diagnosis between 1976 and 1978. Family interviews were possible for 95% (n = 205) of all non-polyposis probands and verification of vital status or medical history was obtained for 98% of 1811 first degree relatives. The prevalence of cancer family syndrome was between 1 and 2%, a figure some fivefold less than that estimated elsewhere. A proximal tumour excess was not characteristic of the ascertained families. These results may have implications for the identification of susceptible people if screening for high risk groups is considered a worthwhile option for reducing colorectal cancer mortality in the United Kingdom.

Histologic characteristics and outcome of familial non-polyposis colorectal carcinoma

Familial cases of non-polyposis colorectal cancer have recently attracted much interest. Little is known about the characteristic histology or natural history of disease in such cases. Our aim was to determine, through a population-based study, whether mucin-secreting tumours were associated with a positive family history and whether 'familiality' was an independent prognostic variable. All patients under 55 years of age with histologically verified colorectal cancer in Northern Ireland during 1976-78 were studied. The family history was validated in 95% of all non-polyposis cases (n = 205), and the proband's histologic specimen reviewed in over 99%. Mucin-secreting tumours were significantly associated with a positive family history, but familiality was not found predictive of survival in a multivariate analysis controlling for age, sex, stage, site, symptom duration, differentiation, and histologic type.