Automated subdural hematoma segmentation for traumatic brain injured (TBI) patients

Abstract

Traumatic brain injury is a serious public health problem in the U.S. contributing to a large portion of permanent disability. However, its early management and treatment could limit the impact of the injury, save lives and reduce the burden of cost for patients as well as healthcare systems. Subdural hematoma is one of the most common types of TBI, which its visual detection and quantitative evaluation are time consuming and prone to error. In this study, we propose a fully auto-mated machine learning based approach for 3D segmentation of convexity subdural hematomas. Textural, statistical and geometrical features of sample points from intracranial region are extracted based on head Computed Tomography (CT) images. Then, a tree bagger classifier is implemented to classify each pixel as hematoma or no-hematoma. Our method yields sensitivity, specificity and area under the receiver operating curve (AUC) of 85:02%, 73:74% and 0:87 respectively.

Fully automated detection and segmentation of meningiomas using deep learning on routine multiparametric MRI

“…Overlap measured by Dice coefficients was high for TTV and for T1CE TV (0.81 and 0.78, respectively). These results are comparable with other recently published studies in brain lesion segmentation using deep learning [14,15,[28][29][30][31] and general segmentation accuracies accounting for intra-and inter-reader variabilities [14,19]. Numerous deep convolutional neuronal networks with different technical specifications have been applied and tested for brain tumour segmentation [13,26].…”

Artificial intelligence and machine learning in emergency medicine

“…The availability of large labelled image datasets makes image analysis a fertile research domain. Deep learning (DL) techniques have been shown to detect pneumonia on chest X‐ray accurately, achieve 3D segmentation of subdural haematomas on brain computed tomography (CT), and assess risk of cerebral aneurysm rupture and score CTs of patients with suspected acute ischaemic stroke as accurately as stroke specialists . DL applied to brain magnetic resonance imaging has been used to distinguish patients with a first episode psychosis from controls, and predict lifetime alcohol consumption .…”

“…Deep learning (DL) techniques have been shown to detect pneumonia on chest X-ray accurately, achieve 3D segmentation of subdural haematomas on brain computed tomography (CT), and assess risk of cerebral aneurysm rupture and score CTs of patients with suspected acute ischaemic stroke as accurately as stroke specialists. [10][11][12][13] DL applied to brain magnetic resonance imaging has been used to distinguish patients with a first episode psychosis from controls, and predict lifetime alcohol consumption. 14,15 DL has also been applied to ultrasound (USG); demonstrating high accuracy in detecting abdominal free fluid on FAST (focused assessment with sonography for trauma) scans, classifying abdominal USG images and providing automated analysis of ejection fraction on echocardiogram.…”

Automated Segmentation and Severity Analysis of Subdural Hematoma for Patients with Traumatic Brain Injuries

“…To the best of our knowledge, there is no prior published work, except our preliminary study [ 8 ], that describes automated methods to detect and segment the different types of SDH. However, there are several techniques developed for acute hematoma detection/segmentation, which are reviewed here.…”

“…In our preliminary work [ 8 ], we reported on 35 CT scans that were included in the current study. The current study expands on this by including a much larger sample size, substantially improving the segmentation algorithm by integrating domain knowledge into data-driven deep models, performing further analyses, and comparing performance with a human benchmark.…”