Prospective external validation of IOTA methods for classifying adnexal masses and retrospective assessment of two‐step strategy using benign descriptors and ADNEX model: Portuguese multicenter study

Abstract

ObjectiveThis study aimed to externally and prospectively validate the International Ovarian Tumor Analysis (IOTA) Simple Rules (SRs), Logistic Regression model 2 (LR2) and Assessment of Different NEoplasias in the adneXa (ADNEX) in a Portuguese population, comparing them with operator subjective assessment (SA), Risk‐of‐Malignancy Index (RMI), as well as with each other. This study also aimed to retrospectively validate IOTA two‐step strategy, using modified benign descriptors (MBDs) followed by the application of ADNEX in cases where MBDs were not applicable (MBDs + ADNEX).MethodsIn this multicenter diagnostic accuracy study, conducted between January 2016 and December 2021, three tertiary referral centers prospectively included consecutive patients with ultrasound diagnosis of at least one adnexal tumor who underwent surgery. All ultrasound assessments were performed by level II or III sonologists with IOTA certification. Patient clinical data and serum cancer antigen (CA125) levels were collected from the hospital databases. Each adnexal mass was classified as benign or malignant using SA, RMI, IOTA SRs, LR2 and ADNEX (with and without CA125). The reference standard was histopathological diagnosis. In the second phase, all adnexal tumors were retrospectively classified using the two‐step strategy (MBDs + ADNEX). The sensitivity, specificity, positive (PPV) and negative predictive value (NPV), positive (LR+) and negative likelihood ratio (LR‐) as well as overall accuracy were determined for SA, RMI, IOTA SRs, LR2, ADNEX and two‐step strategy (MBDs + ADNEX). Receiver‐operator characteristic curves were constructed and corresponding areas under the curve (AUC) determined for RMI, LR2 and ADNEX and two‐step strategy (MBDs + ADNEX). The ADNEX calibration plots were constructed and estimated by LOESS smoother.ResultsOf the 571 included patients, 428 had benign disease, 42 borderline ovarian tumors, 93 primary invasive adnexal cancers and 8 metastatic tumors in adnexa (malignancy prevalence: 25.0%). The operator SA had an overall sensitivity of 97.9% and a specificity of 83.6% for distinguishing between benign and malignant lesions. RMI showed high specificity (95.6%) but very low sensitivity (58.7%), with an AUC of 0.913. The IOTA SRs were applicable in 80.0% of patients, with a sensitivity of 94.8% and a specificity of 98.6%. LR2 revealed a sensitivity of 84.6%, a specificity of 86.9% and an AUC of 0.939, at the malignancy risk cut‐off of 10%. At the same cut‐off, ADNEX with and without CA125 had a sensitivity of 95.8% and 98.6%, respectively, and a specificity of 82.5% and 79.7%, respectively. The AUC of ADNEX with vs. without CA125 was 0.962 vs. 0.960. The ADNEX model provided heterogeneous results in distinguishing between benign and different subtypes of malignancy, with the highest AUC (0.991) for discriminating benign masses from primary adnexal cancer stage II‐IV, and the lowest AUC (0.696) for distinguishing primary adnexal cancer stage I and metastatic lesion in adnexa. The ADNEX calibration plots suggested an underestimation of the predicted risk in relation with the observed proportion of malignancies. The MBDs were applicable in 26.3% of cases (150/571 tumors, none of which were malignant). Similar to the ADNEX model applied in all patients, the two‐step strategy using ADNEX in the second step only, with and without CA125, had an AUC of 0.964 and 0.961, respectively.ConclusionsOur results showed a good to excellent performance of the IOTA methods in the studied Portuguese population, outperforming RMI. ADNEX was superior in accuracy, but interpretation of its ability to distinguish malignant subtypes was fundamentally limited not only by sample size but also by large differences in the prevalence of tumor subtypes. The IOTA MBDs have been shown to be reliable in identifying benign disease. The two‐step strategy based on the application of MBDs, followed by the ADNEX model if MBDs are not applicable, has proven to be suitable for daily practice circumventing the need to use electronic support in all patients.This article is protected by copyright. All rights reserved.