Criteria to interpret cancer biomarker increments crossing the recommended cut-off compared in a simulation model focusing on false positive signals and tumour detection time.

ABSTRACT

BACKGROUND: Several criteria have been proposed to interpret increments in serological cancer biomarker concentrations starting from low baseline concentrations crossing the cut-off. None of the criteria have been compared for their ability to signal tumour growth when ≤2% false positive results are accepted. METHODS: The cancer biomarker Tissue Polypeptide Antigen was used as an example. Seven criteria to interpret increments in concentrations were investigated by computer simulations. Firstly, for each criterion, we identified a baseline concentration stratified for three levels of biological variation providing ≤2% false positive signals of tumour growth during one year of monitoring. Secondly, combining the steady state concentrations with rates of marker increase during tumour growth allowed calculation of the lengths of tumour detection times for each criterion. RESULTS: The number of false positive marker signals depended on the baseline concentration, the magnitude of biological variation, and the magnitude of the required increment defined in the criterion. The lengths of the tumour detection times also depended on the rates of marker increase. CONCLUSIONS: The results suggest that different types of criteria should be used within different intervals of below cut-off level concentrations if the rate of false positive signals of marker increments should be kept ≤2%.