Comparative diagnostic performance of the early-stage ovarian malignancy score versus other risk prediction models in early-stage ovarian cancer: A Vietnamese prospective cohort study.

Objective: This study compares the diagnostic performance of the early-stage ovarian malignancy (EOM) score against other risk prediction models for identifying early-stage ovarian cancer.

Methods: This prospective cohort study involved 925 cases from the obstetrics and gynecology departments of two tertiary hospitals from May 2018 to December 2023. The data included gynecologic examination and/or ultrasound findings, menopausal status, ultrasonography features, serum CA125, and HE4 values, which were used to calculate the EOM score and compare it with other algorithms. Preoperative predictions were validated against postoperative histopathological data.

Results: In total, 792 cases (85.62%) were benign tumors, 74 cases (8.00%) were identified as early-stage ovarian cancer, and 59 cases (6.38%) were classified as advanced-stage ovarian cancer. With a cut-off of ≥13, the EOM score achieved an area under the curve (AUC) value of 0.908 for distinguishing between cancer and non-cancer, demonstrating sensitivity of 83.46% and specificity of 82.90%. For early-stage cancer, the EOM score had an AUC value of 0.843. The EOM score outperformed the risk of malignancy index, the risk of ovarian malignancy algorithm, CPH-I, CA125, and HE4 (P < 0.05).

Conclusion: The EOM score is a straightforward and effective tool for predicting early-stage ovarian cancer, yielded performance similar to the IOTA Simple Rules combined with CA125.