More articles from Adult Brain
- Ensemble of Convolutional Neural Networks Improves Automated Segmentation of Acute Ischemic Lesions Using Multiparametric Diffusion-Weighted MRI
Convolutional neural networks were trained on combinations of DWI, ADC, and low b-value-weighted images from 116 subjects. The performances of the networks (measured by the Dice score, sensitivity, and precision) were compared with one another and with ensembles of 5 networks. An ensemble of convolutional neural networks trained on DWI, ADC, and low b-value-weighted images produced the most accurate acute infarct segmentation over individual networks. Automated volumes correlated with manually measured volumes for the independent cohort.
- Predicting Motor Outcome in Acute Intracerebral Hemorrhage
The authors prospectively studied patients with motor deficits secondary to primary intracerebral hemorrhage within the first 12 hours of symptom onset. Patients underwent multimodal MR imaging including DTI. Intracerebral hemorrhage, perihematomal edema location and volume, and corticospinal tract involvement were assessed. The corticospinal tract was considered affected when the tractogram passed through the intracerebral hemorrhage and/or the perihematomal edema. The authors calculated affected corticospinal tract-to-unaffected corticospinal tract ratios for fractional anisotropy, mean diffusivity, and axial and radial diffusivities. Significant independent predictors of motor outcome were NIHSS and modified NIHSS at admission, posterior limb of the internal capsule involvement by intracerebral hemorrhage at admission, intracerebral hemorrhage volume at admission, 72-hour NIHSS, and 72-hour modified NIHSS. The sensitivity, specificity, and positive and negative predictive values for poor motor outcome at 3 months by a combined modified NIHSS of >6 and posterior limb of the internal capsule involvement in the first 12 hours from symptom onset were 84%, 79%, 65%, and 92%, respectively.