94 For this reason, a non-invasive, internally validated diagnostic prediction model was developed in the Netherlands (204). This model predicts the histopathological diagnosis of adenomyosis based on a combination of clinical characteristics and MRI parameters prior to hysterectomy. The prediction model describes five clinical characteristics: age at MRI, BMI, history of curettage, dysmenorrhea and hypermenorrhea. Additionally, four primary MRI parameters were included in the model: mean junctional zone (JZ) thickness, JZ Differential ³ 5 mm, Junctional Zone to Myometrium Ratio (JZ/MYO) ³ 0.40, and the presence of High Signal Intensity (HSI) foci. Presence of HSI foci was the strongest significant predictor for adenomyosis in this model. The performance of this model showed an Area Under the Curve (AUC) after internal validation of 0.776. This internally validated model can be useful as a prediction tool in patients with suspected adenomyosis and thereafter optimize the management of adenomyosis. The management of the disease depends on various aspects, such as comorbidities, age, impact of the complaints on daily life. Therefore, this tool can assist in shared decision making for both patients and clinicians in this management. The aim of this study was to externally validate this prediction model by Rees et al., so that this model can be clinically implemented for diagnosis of adenomyosis in the general population. Study Objectives The primary objective of this study was hence to perform an external validation of the multivariate diagnostic tool previously constructed by Rees et al. to predict likelihood of histopathological adenomyosis diagnosis based on MRI.
RkJQdWJsaXNoZXIy MTk4NDMw