Tumour markers and screening for gastrointestinal cancer: a follow up study in Finland

Screening for gastrointestinal cancer is based mainly on a barium contrast x ray method and on identification of occult blood in stools. The methods are relatively expensive, not always acceptable to the participants, and there is only limited evidence of their effectiveness in reducing the mortality from gastrointestinal cancer.

To investigate the validity of several tumour markers as a screening test for stomach cancer and for colorectal cancer.

A registry of 36,265 serum samples drawn during 1968-72 was linked to the cancer registry. Follow up was during 1968-80 when 94 stomach cancers and 95 colorectal cancers were identified. One to two matched case-control design was applied, and the concentrations of CEA, CA 19-9, CA 50, and TATI were assessed.

The mean values of the markers between the cases and the controls were almost the same for the total material. Case-control differences were found between the 11 sets with an interval of less than one year between drawing the sample and diagnosis of the cancer. The highest validity was found in CEA for colorectal cancer (specificity 91%; sensitivity 64%) and in CA 19-9 for stomach cancer (specificity 74%; sensitivity 73%).

CEA, CA 19-9, CA 50, or TATI are not valid screening tests. Case-control differences were found with a potential one year screening interval, but they were not large enough for sufficient validity.