Automatic Quantification of Computed Tomography Features in Acute Traumatic Brain Injury.

Saurabh Jain,Vasilis Terzopoulos,Diana Maria Sima,Guido Wilms,Thijs Vande Vyvere,Jan Verheyden,Eloy Roura,Andrew Maas

doi:10.1089/neu.2018.6183

Abstract

Traumatic brain injury is a complex and diverse medical condition with a high frequency of intracranial abnormalities. These can typically be visualized on a computed tomography (CT) scan, which provides important information for further patient management, such as the need for operative intervention. In order to quantify the extent of acute intracranial lesions and associated secondary injuries, such as midline shift and cisternal compression, visual assessment of CT images has limitations, including observer variability and lack of quantitative interpretation. Automated image analysis can quantify the extent of intracranial abnormalities and provide added value in routine clinical practice. In this article, we present icobrain, a fully automated method that reliably computes acute intracranial lesions volume based on deep learning, cistern volume, and midline shift on the noncontrast CT image of a patient. The accuracy of our method is evaluated on a subset of the multi-center data set from the CENTER-TBI (Collaborative European Neurotrauma Effectiveness Research in Traumatic Brain Injury) study for which expert annotations were used as a reference. Median volume differences between expert assessments and icobrain are 0.07 mL for acute intracranial lesions and −0.01 mL for cistern segmentation. Correlation between expert assessments and icobrain is 0.91 for volume of acute intracranial lesions and 0.94 for volume of the cisterns. For midline shift computations, median error is −0.22 mm, with a correlation of 0.93 with expert assessments.

Highlights

Traumatic Brain injury (TBI) is a complex and often poorly understood disease process that is defined as an alteration in brain function, or other evidence of brain pathology, caused by an external force
Three distinctive subcohorts of the CENTERTBI data set are considered for evaluating acute intracranial lesions segmentation: cistern segmentation and midline shift estimation such that every data set ensures a sufficient variability in terms of TBI severity and imaging characteristics of interest
The median volume difference between icobrain and the expert reference segmentations was 0.07 mL, whereas the median value of absolute volume difference was 8.83 mL. icobrain acute intracranial lesions volumes were well correlated to the expert reference volumes with an intraclass correlation coefficient (ICC) of 0.91

Summary

Introduction

Traumatic Brain injury (TBI) is a complex and often poorly understood disease process that is defined as an alteration in brain function, or other evidence of brain pathology, caused by an external force. One of the most important purposes of imaging in the acute phase after injury is to identify the presence of large extra- or intracerebral space-occupying lesions that are in need of urgent neurosurgical evacuation (e.g., subdural hematomas, epidural hematomas, contusions, or intracerebral hematomas). The volume of these lesions and associated secondary features (a midline shift [MLS] greater than 5 mm, cisternal compression, etc.) are important guiding factors for surgical and medical management of raised intracranial pressure.[5,6] In addition, some of these variables are important for outcome prediction, which is why they are used in multiple prognostic CT scoring systems and are commonly collected as important imaging variables in large-scale clinical TBI trials.[7,8,9,10,11,12] Manual segmentation of these lesions is time-consuming and suffers from intra- and interobserver variability.[13,14,15] A fully automated method could increase the reliability and consistency of volume estimations and quantification of associated factors (i.e., MLS and cisternal compression)

Objectives

Methods

Results

Conclusion