Establishment and validation of reference intervals for tumor markers (AFP, CEA, CA19-9, CA15-3, CA125, PSA, HE4, Cyfra 21-1, and ProGRP) in primary care centers in Korea: A cross-sectional retrospective study.

ABSTRACT

Background and Aims: The reference interval (RI) for a tumor marker may vary between populations, detection systems, and the methods used to obtain their values. The aims of this study were to establish age- and sex-specific RIs for the following nine common tumor markers and to validate the established RIs in Korean adults alpha fetoprotein (AFP), carcinoembryonic antigen (CEA), cancer antigen (CA) 19-9, CA15-3, CA125, Human epididymis protein 4 (HE4), total prostate specific antigen, cytokeratin fragment (Cyfra) 21-1, and progastrin-releasing peptide (ProGRP). Methods: This cross-sectional study consecutively selected 214,159 individuals (aged 18-98 years) who underwent health checkups at 16 health-promotion centers in 13 Korean cities. Finally, 62,752 examinees were used to establish the RIs after removing outliers. RIs were established using an indirect method according to the CLSI EP28-A3C guideline. The established RIs were validated by calculating the proportion of individuals outside each RI. Results: Sex-related differences were observed for AFP, CEA, CA19-9, Cyfra 21-1, and ProGRP (p < 0.05) AFP, CEA and Cyfra 21-1 were higher in males, and CA19-9 and proGRP were higher in females. Most of the tumor markers except CA15-3 and CA125 increased with age CA125 decreased at ≥50 years of age (p < 0.05), while CA15-3 did not vary with age. Less than 5% of subjects were outside all RIs (the 2.5th and 97.5th percentiles) established in the present study. Meanwhile, less than 3% of the healthy reference subjects fell outside the current and manufacturers' RIs of all tumor markers except Cyfra 21-1. Conclusion: This study has determined age- and sex-specific RIs for nine common tumor markers in the healthy Korean population, which could be useful for clinicians making clinical decisions and assessments.