Abstract BACKGROUND The expression of RWDD3 is closely related to the prognosis of acromegaly. Therefore, this study aimed to investigate a radiomics method based on MRI to noninvasively evaluate RWDD3 expression in acromegaly. MATERIAL AND METHODS 132 patients with acromegaly were enrolled and divided into primary (n=88) and validation cohorts (n=44) according. The expression of RWDD3 was determined by immunohistochemistry. Radiomic features were extracted from the MR images and determined using the ‘Elastic Net’ feature selection algorithm. A radiomic signature was built using a support vector machine. Subsequently, multivariable logistic regression analysis was used to select the most informative clinical features, and a radiomic model, incorporating the radiomic signature and selected clinical features, was constructed and used as the final predictive model. The performance of this radiomic model was validated using receiver operating characteristics analysis, and its calibration, discriminating ability, and clinical usefulness were assessed. RESULTS The radiomic signature, which was constructed with radiomic features selected using the primary cohort, showed a favorable discriminatory ability in the validation cohort. The radiomic model incorporating the radiomic signature and three selected clinical features showed good discrimination abilities and calibration, with an area under the curve (AUC) of 0.89 for the primary cohort and 0.84 for the validation cohort. The radiomic model better estimated the treatment responses of patients with acromegaly than did the clinical features. Decision curve analysis showed the radiomic model was clinically useful. CONCLUSION This radiomic model could aid neurosurgeons in the prediction of RWDD3 expression in patients with acromegaly, and could contribute to predicting of patient prognoses.