Clinical NeuroscienceVoxel-based Gaussian naïve Bayes classification of ischemic stroke lesions in individual T1-weighted MRI scans

•We present a novel method for lesion delineation in individual T1 MRI scans.•We compared our method with manual delineation for a large group of stroke patients.•Our method reliably predicted lesion extents and volumes.•Our method identified lesion effects that pose challenges for manual delineation.•Our method can be used for lesion-symptom mapping and clinical volume estimation.

BackgroundManual lesion delineation by an expert is the standard for lesion identification in MRI scans, but it is time-consuming and can introduce subjective bias. Alternative methods often require multi-modal MRI data, user interaction, scans from a control population, and/or arbitrary statistical thresholding.New methodWe present an approach for automatically identifying stroke lesions in individual T1-weighted MRI scans using naïve Bayes classification. Probabilistic tissue segmentation and image algebra were used to create feature maps encoding information about missing and abnormal tissue. Leave-one-case-out training and cross-validation was used to obtain out-of-sample predictions for each of 30 cases with left hemisphere stroke lesions.ResultsOur method correctly predicted lesion locations for 30/30 un-trained cases. Post-processing with smoothing (8 mm FWHM) and cluster-extent thresholding (100 voxels) was found to improve performance.Comparison with existing methodQuantitative evaluations of post-processed out-of-sample predictions on 30 cases revealed high spatial overlap (mean Dice similarity coefficient = 0.66) and volume agreement (mean percent volume difference = 28.91; Pearson's r = 0.97) with manual lesion delineations.ConclusionsOur automated approach agrees with manual tracing. It provides an alternative to automated methods that require multi-modal MRI data, additional control scans, or user interaction to achieve optimal performance. Our fully trained classifier has applications in neuroimaging and clinical contexts.