Diagnosis of epithelial ovarian cancer in women with ovarian masses using multiplex immunoassays and krestchmann-surface plasmon resonance

Abstract

Epithelial ovarian cancer (EOC) has a high mortality rate, largely attributable to late detection and limitations in current diagnostic methods for distinguishing between malignant and benign ovarian masses. This study aimed to evaluate the diagnostic performance of multiple serum biomarkers in differentiating malignant from benign ovarian masses, to identify high-performing anchor biomarkers for integration into a multi-biomarker model, and to assess the feasibility of a label-free K-SPR biosensor for rapid and cost-effective point-of-care biomarker detection. A total of 34 serum samples (benign n = 18, malignant n = 16) were analysed using multiplex

immunoassay to measure ten biomarkers (CA125, HE4, IL-6, TNFR2, CCL11, IL- 2Rα, VCAM1, PRL, OPN, and MIF), and receiver operating characteristic (ROC)

analysis was performed to evaluate their diagnostic performance and develop multi- biomarker panels. Subsequently, a subset of samples (n = 8) was analysed using a

label-free Kretschmann-based surface plasmon resonance (K-SPR) biosensor to assess the linear detection performance of selected biomarkers. The results showed that CA125 (~7.8-fold increase) and HE4 emerged as the strongest biomarkers. HE4 demonstrated 100% specificity but moderate sensitivity (60%) (AUC = 0.951). For CA125, the 60 U/mL cut-off improved sensitivity (93.75%) and accuracy (97.00%) compared to the conventional 35 U/mL threshold (68.18% sensitivity, 79.00% accuracy). OPN and inflammatory markers (IL-6, MIF, TNFR2) showed moderate performance (AUC ≥ 0.70), supporting a complementary role, while CCL11, PRL, VCAM1, and IL-2Rα demonstrated poor diagnostic performance. The CA125 + HE4 + IL-6 panel demonstrated strong and stable performance (AUC = 0.979), while the CA125 + HE4 + IL-6 + MIF panel achieved perfect classification (AUC = 1.000) but exhibited lower stability upon validation, indicating possible overfitting. In K-SPR serum samples, CA125 differentiated benign (<60 U/mL) from malignant (96.1– 716.36 U/mL) samples with strong linearity (R2>0.94). HE4 showed clear discrimination between benign (<46 pM) and malignant (>61.55 pM) samples at a 60 pM threshold, also exhibiting robust linearity (R2 ≈ 0.95). In conclusion, the CA125 + HE4 + IL-6 panel demonstrated the highest reliability for distinguishing ovarian cancer, while the integrated K-SPR biosensor showed strong potential as a diagnostic tool. Overall, combining multiple biomarkers with this biosensor technology may improve EOC detection, offering a faster, more affordable, and simpler alternative to current methods. However, further validation in larger patient cohorts is required before it can be widely implemented in clinical practice.