Abstract
Objectives: CA19-9 elevation in non-gastrointestinal tumor patients may be influenced by various non-tumor factors, which poses challenges for clinical diagnosis. This study aims to assess the consistency between initial elevated CA19-9 levels detected by the ARCHITECT/Alinity i system (Abbott Diagnostics) and subsequent retesting using the Elecsys CA19-9 assay (Roche Diagnostics) in 5372 non-gastrointestinal tumor patients, and to explore potential factors contributing to CA19-9 non-specific elevation. Methods: Bland-Altman and Passing-Bablok analyses were used to assess the agreement between the two assays. Nonparametric Spearman and Pearson's chi-square tests were used to assess the correlation between CA19-9 and different clinical comorbidities/antigen concentration strata and to compare the categorization by age/disease, respectively. Results: Bland-Altman and Passing-Bablok regression analyses revealed that the CA19-9 test results from Abbott and Roche platforms show significant systematic bias and weak correlation, making the two methods not directly interchangeable. After excluding common confounders, the study focused on heterophilic antibodies (HAs) as target. Blood samples were treated with a commercial blocking agent demonstrated alignment with baseline Elecsys CA19-9 results but differed significantly from initial ARCHITECT/Alinity i measurements. Furthermore, non-specific CA19-9 elevation was also associated with comorbidities such as diabetes mellitus, pulmonary infections, breast nodules, uterine leiomyoma, and its incidence increased with age. Conclusions: The study highlights the need to consider potential interferences and underlying disorders when results conflict with clinical diagnoses. Method-specific validation and comprehensive clinical correlation are crucial for accurate interpretation of CA19-9 levels to prevent misdiagnosis and ensure appropriate patient management.