Future Unruptured Intracranial Aneurysm Growth Prediction Using Mesh Convolutional Neural Networks

The growth of unruptured intracranial aneurysms (UIAs) is a predictor rupture. Therefore, for further imaging surveillance and treatment planning, it important to be able predict if an UIA likely grow based on initial baseline Time-of-Flight MRA (TOF-MRA). It known that the size shape UIAs are predictors aneurysm and/or We perform feasibility study using mesh convolutional neural network future prediction from TOF-MRAs. include 151 TOF-MRAs, with 169 where 49 were classified as growing 120 stable, clinical definition (>1 mm increase in follow-up scan). segmented TOF-MRAs meshes automatically generated. investigate input both only region-of-interest (ROI) including surrounding parent vessels. develop classification model will or remain stable. consisted additional novel edge features index curvedness which describe surface topology. was investigated mid-point co-ordinates influenced performance. highest AUC (63.8%) co-ordinate (average F1 score = 62.3%, accuracy 66.9%, sensitivity 57.3%, specificity 70.8%). present promising results.