Accuracy of an nnUNet neural network for the automatic segmentation of intracranial aneurysms, their parent vessels and major cerebral arteries from magnetic resonance imaging-Time of flight (MRI-TOF).

Abstract

To develop a new machine-learning algorithm for fully automatic identification of cerebral arteries and intracranial aneurysms (IAs) based on a manually segmented magnetic resonance imaging with time-of-flight sequences (MRITOF) dataset. In this retrospective single-center study, 62 MRI-TOF scans of a total of 73 untreated unruptured IAs were manually color-labelled in 21 classes. A nnUNet architecture was trained on MRI-TOF images. The performance of the automatic segmentation was compared with the manual segmentation using Dice Similarity Coefficient (DSC), Centerline Dice (ClDice) and 95th percentile Hausdorff Distance (HD95). Sensitivity was computed for aneurysm detection. Across all 21 classes, the median DSC was 0.86 [95CI: 0.81, 0.89], the median ClDice 0.91 [0.85, 0.94] and the median HD95 was 2.9 [1.0, 14.9] mm. Sensitivity of the model for aneurysms detection was 0.8. For this class specifically, a median DSC of 0.88 [0.13, 0.92], median ClDice of 0.89 [0.06, 1.0] and median HD95 of 1.8 [0.58, 81] mm was achieved. The volume of the labelled anatomical structure was the most relevant determinant of accuracy in this model. Median time to predict was 130.6 [60.9, 284.1] seconds. The nnUNet MRI-TOF based algorithm provided a fast and adequate automatic extraction of unruptured intracranial aneurysms, their parent vessels and the most relevant cerebral arteries. Future steps involve the expansion of the training set with the inclusion of more MRI-TOF studies with and without IAs and its incorporation in 3D imaging viewers and treatment prediction. IA = Intracranial Aneurysm; MRI-TOF＝ Magnetic Resonance Imaging - Time of Flight; DSC = Dice-Sørenson Coefficient; ClDice = Centerline Dice; HD95 = 95th Percentile Hausdorff Distance.