Comparison of diagnostic efficiency between IOTA LR2 model and doctors' experiences.

Abstract

International Ovarian Tumor Analysis (IOTA) working group proposed a logistic regression (IOTA LR2) model. It is served as a risk prediction model for benign and malignant adnexal tumors. This study aims to compare the diagnostic efficiency between the IOTA LR2 model and doctors' subjective assessment on diagnosing benign and malignant adnexal mass. The ultrasonographic images of 616 adnexal masses were retrospectively analyzed by the senior doctors' group and the junior doctors' group using the IOTA LR2 model and subjective assessment. The postoperative pathological diagnosis was used as the gold standard to compare the diagnostic efficiency of the 2 methods. The area under the curves of subjective assessment and IOTA LR2 model for diagnosing malignant adnexal masses were 0.86 and 0.90 for the senior doctors' group and 0.79 and 0.88 for the junior doctors' group, respectively. The sensitivity and specificity of subjective assessment for diagnosing the malignant adnexal masses were 81.0% and 91.3% for the senior doctors' group and 70.1% and 88.7% for the junior doctors' group, respectively. The sensitivity and specificity of the IOTA LR2 model for diagnosing the malignant adnexal masses were 79.6% and 88.1% for the senior doctors' group, and 79.6% and 81.7% for the junior doctors' group, respectively.There were no significant difference in the sensitivities between the senior doctors' group and junior doctors' group using the IOTA LR2 model and the senior doctors' group using subjective assessment (both P>0.05). The diagnostic efficiency of the IOTA LR2 model is equal to the senior doctors' experiences. This model can help junior doctors to reduce the missed diagnosis of malignant adnexal masses.