Editor’s Choice: Ensemble of Convolutional Neural Networks Improves Automated Segmentation of Acute Ischemic Lesions Using Multiparametric Diffusion-Weighted MRI

Editor’s Choice

The authors analyzed the accuracy of gadolinium SWI for detecting the imaging evidence of active inflammation on MS plaques when a BBB dysfunction was demonstrated by a focal gadolinium-enhanced lesion and compared this technique with gadolinium-enhanced T1 spin-echo and T1 spin-echo with magnetization transfer contrast sequences. Differences in BBB dysfunction were evident in the 103 patients among gadolinium SWI, gadolinium T1 spin-echo, and gadolinium T1 magnetization transfer contrast. Gadolinium T1 magnetization transfer contrast demonstrated the highest number of active demyelinating plaques. Gadolinium SWI was highly correlated with gadolinium T1 magnetization transfer contrast in depicting acute demyelinating plaques and these techniques provided better performance compared with gadolinium T1 spin-echo.

 

Abstract

BACKGROUND AND PURPOSE

Sample segmentation results of the ensemble of DWI+ADC+LOWB (blue regions) on sample subjects along with manual outlines (red outlines).

Accurate automated infarct segmentation is needed for acute ischemic stroke studies relying on infarct volumes as an imaging phenotype or biomarker that require large numbers of subjects. This study investigated whether an ensemble of convolutional neural networks trained on multiparametric DWI maps outperforms single networks trained on solo DWI parametric maps.

MATERIALS AND METHODS

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. To assess the generalizability of the approach, we applied the best-performing model to an independent Evaluation Cohort of 151 subjects. Agreement between manual and automated segmentations for identifying patients with large lesion volumes was calculated across multiple thresholds (21, 31, 51, and 70 cm3).

RESULTS

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 (P < .001). Automated volumes correlated with manually measured volumes (Spearman ρ = 0.91, P < .001) for the independent cohort. For the task of identifying patients with large lesion volumes, agreement between manual outlines and automated outlines was high (Cohen κ, 0.86–0.90; P < .001).

CONCLUSIONS

Acute infarcts are more accurately segmented using ensembles of convolutional neural networks trained with multiparametric maps than by using a single model trained with a solo map. Automated lesion segmentation has high agreement with manual techniques for identifying patients with large lesion volumes.

Read this article: http://bit.ly/2JX4mar

Editor’s Choice: Ensemble of Convolutional Neural Networks Improves Automated Segmentation of Acute Ischemic Lesions Using Multiparametric Diffusion-Weighted MRI