Diagnostic accuracy of the risk of malignancy index for adnexal masses in women over 35 years: a prospective cross-sectional study at a tertiary care centre in Punjab, India.
Student's Journal of Health Research Africa · 2026-09-30 · 7 (3) · 10
DOI: 10.51168/qqn8qs93 ↗Abstract
Background: The risk of malignancy index (RMI) is widely used to decide which adnexal masses should be referred to a gynaecological oncologist, but few Indian studies report which tumours the index misses. This study estimated the accuracy of RMI at the conventional cut-off of 200 against histopathology, compared it with its individual components, and examined misclassified cases by histological type and menopausal status. Methodology: This prospective cross-sectional study enrolled 80 consecutive women aged over 35 years with an adnexal mass scheduled for surgery at a tertiary care teaching hospital in Punjab, India, from 23 April 2021 to 22 October 2022. RMI was calculated as U × M × CA-125, with the ultrasound score (U) and menopausal score (M) each 1 or 3. Histopathology was the reference standard. Results: Forty-seven of 80 masses (58.8%) were malignant. At a cut-off of 200, RMI had a sensitivity of 87.2% (95% CI 74.3–95.2), specificity of 87.9% (71.8–96.6), positive likelihood ratio 7.2 and negative likelihood ratio 0.15. The area under the ROC curve was 0.94 (0.88–0.99), higher than for CA-125 (0.84; P = .002), ultrasound score (0.85; P = .043) or menopausal status (0.83; P = .008) alone. Five of six false negatives were mucinous carcinomas: RMI detected 2 of 7 mucinous carcinomas against 39 of 40 other malignancies (P < .001). Specificity in postmenopausal women was 4 of 7. Conclusion: RMI at 200 performed as reported in other populations. A lower in-sample optimum near 160 rested on three mucinous carcinomas and is too unstable to replace 200. A low RMI should not overrule a malignant-appearing ultrasound, particularly when a mucinous tumour is possible. Recommendation: Multicentre studies that include women managed without surgery, document blinding, and compare RMI with International Ovarian Tumor Analysis (IOTA) models in the same patients are needed.
EUVIMED