Prediction of hemifacial spasm (HFS) re-appearing phenomenon after surgery in patients with HFS using DSC perfusion MRI

Motivation: DSC MRI may provide clues to predict hemifacial spasm (HFS) re-appearing phenomenon after surgery (HFSrapas) after microvascular decompression (MVD). Goal(s): To predict HFSrapas with DSC parameter indices using a machine learning analysisApproach: Sixty patients who underwent MVD for HFS were enrolled. DSC parameters were used to predict HFSrapas using a ROC curve and machine learning methods. Results: The rCBF value was significantly decreased in the reappearing group. The extraction fraction parameter was best predicted by the Navie Bayes (NB) model. Impact: DSC perfusion MRI is a useful tool to predict HFS recurrence before intra-operation and helps neurosurgeons anticipate possible problems during MVD surgery.