EP23.05: ADNEX versus LR2: a meta‐analysis comparing the performance of two quantitative tests for differentiating adnexal masses

To compare the sensitivities and specificities of two quantitative ultrasound models, IOTA ADNEX and logistic regression 2 (LR2) in differentiating between benign and malignant pathology in known adnexal masses. This meta-analysis used the MEDLINE database to find primary articles which compared various models of evaluating adnexal masses for malignancy. To account for recent advances in ultrasound technology, the search was limited to articles published between 2016 and 2024. 204 articles were investigated by one reviewer for adnexal masses which were assessed using LR2 or ADNEX without CA-125 and confirmed by postoperative pathology. Across the 21 articles which met inclusion criteria, the results of assessing 7,396 adnexal masses were extracted and compared using ultrasound classification as test outcomes and postoperative pathology as true outcomes. The cutoff value for malignancy using ADNEX and LR2 was 0.10 for all masses. The sensitivity of ADNEX was found to be 91.3%, while LR2 demonstrated a sensitivity of 84%. The difference in sensitivity between the two criteria was statistically significant (p < 0.00001), indicating that ADNEX outperformed LR2 in identifying malignancy correctly. Regarding specificity, ADNEX exhibited a specificity of 84.1%, while LR2 had a specificity of 84.5%. The p-value for this comparison was found to be 0.64, indicating no statistically significant difference in specificity between ADNEX and LR2. Our findings suggest that while both ADNEX and LR2 show comparable specificity, ADNEX demonstrates significantly higher sensitivity in detecting adnexal malignancy. These results emphasise the potential superiority of ADNEX over LR2 as an ultrasound criterion for the assessment of adnexal masses. We propose that more research should be conducted to determine more fully whether ADNEX should replace LR2 as a gold standard criterion.