RU-Net: skull stripping in rat brain MR images after ischemic stroke with rat U-Net

Abstract

Background: Experimental ischemic stroke models play a fundamental role in interpreting the mechanism of cerebral ischemia and appraising the development of pathological extent. An accurate and automatic skull stripping tool for rat brain image volumes with magnetic resonance imaging (MRI) are crucial in experimental stroke analysis. Due to the deficiency of reliable rat brain segmentation methods and motivated by the demand for preclinical studies, this paper develops a new skull stripping algorithm to extract the rat brain region in MR images after stroke, which is named Rat U-Net (RU-Net).

Methods: Based on a U-shape like deep learning architecture, the proposed framework integrates batch normalization with the residual network to achieve efficient end-to-end segmentation. A pooling index transmission mechanism between the encoder and decoder is exploited to reinforce the spatial correlation. Two different modalities of diffusion-weighted imaging (DWI) and T2-weighted MRI (T2WI) corresponding to two in-house datasets with each consisting of 55 subjects were employed to evaluate the performance of the proposed RU-Net.

Results: Extensive experiments indicated great segmentation accuracy across diversified rat brain MR images. It was suggested that our rat skull stripping network outperformed several state-of-the-art methods and achieved the highest average Dice scores of 98.04% (p < 0.001) and 97.67% (p < 0.001) in the DWI and T2WI image datasets, respectively.

Conclusion: The proposed RU-Net is believed to be potential for advancing preclinical stroke investigation and providing an efficient tool for pathological rat brain image extraction, where accurate segmentation of the rat brain region is fundamental.

Keywords: Brain segmentation; Deep learning; Ischemic stroke; MRI; Skull stripping; U-Net.

Fig. 1

Illustration of the proposed RU-Net architecture

Fig. 2

Plots of the accuracy and loss functions using the RU-Net in the training and validation datasets. Top row: DWI subjects. Bottom row: T2WI subjects

Fig. 3

Illustration of DWI (Subject 39) skull stripping results using the proposed RU-Net framework. Yellow: Prediction. Red: GT

Fig. 4

Performance analyses of DWI skull stripping results based on five-fold cross validation

Fig. 5

Visual comparison of DWI skull stripping results using different methods. Top row: slices 7 and 8 of Subject 10. Bottom row: slices 9 and 10 of Subject 21

Fig. 6

Visual comparison of DWI skull stripping results in 3-D view using different methods. Blue: . Red: . Top row: Subject 16. Bottom row: Subject 40

Fig. 7

Illustration of T2WI (Subject 31) skull stripping results using the proposed RU-Net framework. Yellow: Prediction. Red: GT

Fig. 8

Performance analyses of T2WI skull stripping results based on five-fold cross validation

Fig. 9

Visual comparison of T2WI skull stripping results using different methods. Top row: slices 6 and 7 of Subject 9. Bottom row: slices 8 and 9 of Subject 33

Fig. 10

Visual comparison of T2WI skull stripping results in 3-D view using different methods. Blue: . Red: . Top row: Subject 3. Bottom row: Subject 37

Fig. 11

Visual skull stripping results using the original MU-Net (top row) and MedicDeepLabv3+ (bottom row) without retraining. Left columns: DWI Subjects 10 and 21. Right columns: T2WI Subjects 9 and 33