AUTHOR=Xu Jun , Zhang Rongguo , Zhou Zijian , Wu Chunxue , Gong Qiang , Zhang Huiling , Wu Shuang , Wu Gang , Deng Yufeng , Xia Chen , Ma Jun 

TITLE=Deep Network for the Automatic Segmentation and Quantification of Intracranial Hemorrhage on CT

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 14 - 2020

YEAR=2021

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2020.541817

DOI=10.3389/fnins.2020.541817

ISSN=1662-453X

ABSTRACT=Background: ABC/2 method was usually applied to evaluate intracerebral hemorrhage (ICH) volume on computed tomography (CT), while it might be inaccurate and not applicable to estimating extradural or subdural hemorrhage (EDH, SDH) volume due to their irregular hematoma shapes. This study aimed to evaluate a deep framework optimized for the segmentation and quantification of ICH, EDH and SDH. 
Methods: The training datasets were 3000 images retrospectively collected from a collaborating hospital (Hospital A) and segmented by the Dense U-Net framework. Three experienced radiologists determined the ground truth by marking the pixels as hemorrhage area and we utilized the Dice and intra-class correlation coefficients (ICC) to test the reliability of the ground truth. Moreover, the testing datasets consisted of 211 images (internal test) from Hospital A, and 86 ICH images (external test) from another hospital (Hospital B). In this study, we chose scatter plots, ICC and Pearson correlation coefficient (PCC) with ground truth to evaluate the performance of the deep framework. Furthermore, to validate the effectiveness of the deep framework, we did a comparative analysis of the hemorrhage volume estimation between the deep model and ABC/2 method.
Results: The high Dice (0.89-0.95) and ICC (0.985-0.997) showed the consistency of the manual segmentations among the radiologists and the reliability of the ground truth. For internal test, the Dice coefficients of ICH, EDH and SDH were 0.90 ± 0.06, 0.88 ± 0.12 and 0.82 ± 0.16, respectively. For external test, the segmentation Dice was 0.86 ± 0.09. Comparatively, the ICC and PCC of ICH volume estimations were 0.99 performed by Dense U-Net that overmatched ABC/2 method.
Conclusions: This study revealed the excellent performance of hematoma segmentation and volume evaluation based on Dense U-Net, which indicated our deep framework might contribute to efficiently developing treatment strategies for intracranial hemorrhage in clinic.