111 to all patient suspected of adenomyosis in general. Potentially, our study population consists of patients with more severe complaints who therefore undergo an MRI and hysterectomy. This model could be used in clinical practice to predict the chance of adenomyosis in an individual patient. A notable issue with using the model is that the interpretation of outcome percentages varies by individual. Some patients may find a chance of 40 percent high, while another patient only desires to start a particular treatment at a 70 percent chance for example. Therefore, a cut-off value should be discussed. Future research should perform an external validation in a larger cohort to confirm the model’s generalisability and performance in different settings. Further investigation of the clinical usefulness and impact of the model in daily practice is also needed. Conclusions In conclusion, the developed model showed good to excellent discriminative performance in this external cohort for predicting the adenomyosis diagnosis based on MRI in individual patients. The model could be used in clinical practice and could aid in shared decision-making of the subsequent management of this disease, in conjunction with other tests and clinical information. It may be beneficial to involve gynaecologists in assessing the MRI parameters needed for the prediction model, given the suboptimal diagnostic accuracy of radiology reports in detecting adenomyosis. This could enhance the reliability of the diagnostic process. Larger (prospective) are needed before utilizing this model in daily practice.
RkJQdWJsaXNoZXIy MTk4NDMw