Modified Risk of Ovarian Malignancy Algorithm and Risk of Malignancy Index in Predicting Epithelial Ovarian Cancer in Indonesian Population: A Single-centered Validation Study

Background: Risk of ovarian malignancy algorithm (ROMA) and risk of malignancy index (RMI) are two scoring systems that are commonly used to predict ovarian tumor malignancy. Literature shows different cutoff points for a different population. Objective: This study aims to validate and compare the performance of ROMA and RMI and also validate the cutoff points for Indonesian population. Methods: This is a retrospective study conducted at Dr Cipto Mangunkusumo Hospital (CMH). Medical records of patients with epithelial ovarian cancer who underwent surgery in our institution during 2010–2014 were collected. The diagnostic values of ROMA and RMI were calculated. Results: From the analysis of 213 subjects included in this study, ROMA was statistically better than RMI [AUC (area under the curve) in all groups: 87.00% > 81.30%, p ≤0.001; postmenopausal group: AUC 91.47% > 88.97%, p ≤0.001]. RMI had values of sensitivity: 85.3%, specificity: 66.3%, positive predictive value (PPV): 79.7%, negative predictive value (NPV): 74.3%, positive likelihood ratio (LR): 2.53, negative LR: 0.22, and accuracy: 0.77. ROMA had values of sensitivity: 95.4%, specificity: 32.5%, PPV: 68.9%, NPV: 81.8%, positive LR: 1.41, negative LR: 0.14, and accuracy: 0.71. At the ideal cutoff point (RMI 330, premenopausal ROMA 30.4, and postmenopausal ROMA 53.1), ROMA showed better sensitivity and specificity than RMI (sensitivity of 82.31 vs 74.62%; specificity of 78.31 vs 75.9%). Conclusion: ROMA is better than RMI in predicting epithelial ovarian cancer in Indonesian population. Using the modified cutoff, the specificities of both ROMA and RMI were better than the standard cutoff points.