Predictability of radiosurgical outcome in vestibular schwannomas: the current state of MRI-based machine learning prediction models

Vestibular schwannomas (VS) exhibit highly variable response patterns following stereotactic radiosurgery (SRS), ranging from regression and stability to transient enlargement (TTE) or treatment failure. This unpredictability complicates clinical decision-making. Although artificial intelligence (AI) enables large-scale analysis of medical images for predictive modeling, to date, no robust models for post-SRS VS response have been implemented in clinical practice. In this study, a systematic review is conducted on the current landscape of image-based prediction models for post-SRS VS response.