An evaluation of interpretative criteria for liver scintigraphy in a population with suspected or proven colo-rectal cancer. A prospective, consecutive study.

Radioisotope liver scintigraphy is a method commonly used to reveal metastases. In clinical trials the method is found at least as good as biochemical tests, ultrasonography or computed tomography. However, despite technical progress, the clinical value of the test seems not to have improved. By using three cutoff levels in the scintigraphic interpretation--depending on the degree of suspicion of focal defect(s)--an attempt to improve the clinical value of the test is evaluated in a population consisting of 142 patients strongly suspected of having colo-rectal cancer. The incidence of liver metastases was 0.15. In the diagnostic situation the best scintigraphic cutoff level gives the predictive value of a positive test = 0.71 with the nosological false-negative ratio = 0.52. In the screening situation the best cutoff level gives the predictive value of a negative test = 0.96 with the nosological false-positive ratio = 0.21. The benefit of close cooperation between clinicians and physicians in nuclear medicine in order to define the clinical problem and the scintigraphic cutoff level is accordingly emphasized.