Deep learning radiomic nomogram can distinguish intracranial solitary fibrous tumor from angiomatous meningioma: a multicenter study

Motivation: A novel and noninvasive method for distinguishing intracranial solitary fibrous tumor (ISFT) from angiomatous meningioma (AM) and predicting patient outcomes is urgent. Goal(s): To evaluate the value of a MRI-based deep learning radiomic nomogram (DLRN) in distinguishing ISFT from AM and predicting patient outcomes. Approach: A MRI-based DLRN was developed on training cohort (TC). We then validated it's performance on external validation cohort (EVC). Moreover, we investigated the value of the DLRN in survival analysis. Results: The performance of DLRN was excellent (0.86 [0.84–0.88]) on EVC. Besides, DLRN was significantly associated with the overall survival (OS) of patients (n=273). Impact: The proposed DLRN can potentially provide a noninvasive method for neurosurgeon to offer decision support for developing personalized treatment plans and predicting patient outcomes.