Biomarker-Based Models for Preoperative Assessment of Adnexal Mass: A Multicenter Validation Study

Area of Science:

• Oncology
• Proteomics
• Biomarker Discovery

Background:

• Ovarian cancer (OC) is a highly lethal gynecologic malignancy.
• Accurate non-invasive diagnostic methods are crucial for early detection and improved patient outcomes.

Purpose of the Study:

• To develop and validate novel proteomic-based models for the non-invasive diagnosis of ovarian cancer.
• To compare the performance of new models against existing algorithms like ROMA-50 and CPH-I, and a previous proteomic model.

Main Methods:

• Serum samples from 356 women (279 training, 77 validation) with benign or malignant ovarian tumors were analyzed.
• Bead-based multiplex technology (Luminex®) was used to assess six serum proteins: CA125, HE4, osteopontin (OPN), prolactin, leptin, and macrophage migration inhibitory factor (MIF).
• Proteomic Model 2021, Combined Model 2021 (Proteomic Model 2021 + age), and Full Combined Model 2021 (all six proteins + age) were developed and validated.

Main Results:

• Proteomic Model 2021 achieved a c-index of 0.98 (training) and 0.82 (validation).
• Combined Model 2021 showed c-indices of 0.99 (training) and 0.86 (validation).
• Full Combined Model 2021 yielded c-indices of 0.98 (training) and 0.89 (validation).
• All developed models demonstrated high specificity and sensitivity (>0.9) in postmenopausal women, outperforming ROMA-50.
• Performance was poor in women under 50 or with normal CA125 levels.

Conclusions:

• The developed proteomic models, particularly the Combined Model 2021, show satisfactory diagnostic accuracy for ovarian cancer, especially in postmenopausal women.
• The Combined Model 2021, utilizing four proteins and age, performed comparably to the Full Combined Model 2021.
• CPH-I and Proteomic Model 2017 also showed good diagnostic performance, but no single model was clearly superior across all metrics.