New Analytical Approach for the Alignment of Different {HE}4 Automated Immunometric Systems: An Italian Multicentric Study

Human epididymal secretory protein 4 (HE4) elevation has been studied as a crucial biomarker for malignant gynecological cancer, such us ovarian cancer (OC). However, there are conflicting reports regarding the optimal HE4 cut-off. Thus, the goal of this study was to develop an analytical approach to harmonize HE4 values obtained with different laboratory resources. To this regard, six highly qualified Italian laboratories, using different analytical platforms (Abbott Alinity I, Fujirebio Lumipulse G1200 and G600, Roche Cobas 601 and Abbott Architett), have joined this project. In the first step of our study, a common reference calibration curve (designed through progressive HE4 dilutions) was tested by all members attending the workshop. This first evaluation underlined the presence of analytical bias in different devices. Next, following bias correction, we started to analyze biomarkers values collected in a common database (1509 patients). A two-sided p-value < 0.05 was considered statistically significant. In post-menopausal women stratified between those with malignant gynecological diseases vs. non-malignant gynecological diseases and healthy women, dichotomous HE4 showed a significantly better accuracy than dichotomous Ca125 (AUC 0.81 vs. 0.74, p = 0.001 for age ≤ 60; AUC 0.78 vs. 0.72, p = 0.024 for age > 60). Still, in post-menopausal status, similar results were confirmed in patients with malignant gynecological diseases vs. patients with benign gynecological diseases, both under and over 60 years (AUC 0.79 vs. 0.73, p = 0.006; AUC 0.76 vs. 0.71, p = 0.036, respectively). Interestingly, in pre-menopausal status women over 40 years, HE4 showed a higher accuracy than Ca125 (AUC 0.73 vs. 0.66, p = 0.027), thus opening new perspective for the clinical management of fertile patients with malignant neoplasms, such as ovarian cancer. In summary, this model hinted at a new approach for identifying the optimal cut-off to align data detected with different HE4 diagnostic tools.