Efficiency of composite tests in gastroentestinal cancer

The efficiency of composite tests (liver scintigraphy, serum alkaline phosphatase, and serum carcinoembryonic antigen) in findings or excluding liver metastases preoperatively was evaluated in 185 surgical patients with high probability for gastrointestinal cancer, 142 with colorectal and 43 with gastric disorders. A pathoanatomic verification procedure showed liver metastases in 21 and 7 patients, respectively. For each test two cut-off levels were defined in accordance with the operational purpose of the test: either to diagnose metastases (no false-positive test results) or to exclude metastases (no false-negative test results). Generally, composite tests increased overall efficiency. In the colorectal group 39% of the patients were correctly classified by the combined, triple test; in the gastric group 94% were correctly classified. It is concluded that composite tests seem useful, and the operational approach described may be helpful in decision-making and test evaluation