Automated quantification of cerebral edema following hemispheric infarction

Application of a machine-learning algorithm to evaluate CSF shifts on serial head CTs

Yasheng Chen, Rajat Dhar, Laura Heitsch, Andria Ford, Israel Fernandez-Cadenas, Caty Carrera, Joan Montaner, Weili Lin, Dinggang Shen, Hongyu An, Jin Moo Lee

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Although cerebral edema is a major cause of death and deterioration following hemispheric stroke, there remains no validated biomarker that captures the full spectrum of this critical complication. We recently demonstrated that reduction in intracranial cerebrospinal fluid (CSF) volume (∆ CSF) on serial computed tomography (CT) scans provides an accurate measure of cerebral edema severity, which may aid in early triaging of stroke patients for craniectomy. However, application of such a volumetric approach would be too cumbersome to perform manually on serial scans in a real-world setting. We developed and validated an automated technique for CSF segmentation via integration of random forest (RF) based machine learning with geodesic active contour (GAC) segmentation. The proposed RF + GAC approach was compared to conventional Hounsfield Unit (HU) thresholding and RF segmentation methods using Dice similarity coefficient (DSC) and the correlation of volumetric measurements, with manual delineation serving as the ground truth. CSF spaces were outlined on scans performed at baseline (< 6 h after stroke onset) and early follow-up (FU) (closest to 24 h) in 38 acute ischemic stroke patients. RF performed significantly better than optimized HU thresholding (p < 10− 4 in baseline and p < 10− 5 in FU) and RF + GAC performed significantly better than RF (p < 10− 3 in baseline and p < 10− 5 in FU). Pearson correlation coefficients between the automatically detected ∆ CSF and the ground truth were r = 0.178 (p = 0.285), r = 0.876 (p < 10− 6) and r = 0.879 (p < 10− 6) for thresholding, RF and RF + GAC, respectively, with a slope closer to the line of identity in RF + GAC. When we applied the algorithm trained from images of one stroke center to segment CTs from another center, similar findings held. In conclusion, we have developed and validated an accurate automated approach to segment CSF and calculate its shifts on serial CT scans. This algorithm will allow us to efficiently and accurately measure the evolution of cerebral edema in future studies including large multi-site patient populations.

Original languageEnglish
Pages (from-to)673-680
Number of pages8
JournalNeuroImage: Clinical
Volume12
DOIs
Publication statusPublished - 2016

Fingerprint

Fluid Shifts
Brain Edema
Infarction
Cerebrospinal Fluid
Head
Stroke
Tomography
Machine Learning
Cause of Death
Biomarkers
Population

Keywords

  • Active contour
  • Cerebral edema
  • CSF segmentation
  • Ischemic stroke CT
  • Mass effect
  • Random forest

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Neurology
  • Clinical Neurology
  • Cognitive Neuroscience

Cite this

Automated quantification of cerebral edema following hemispheric infarction : Application of a machine-learning algorithm to evaluate CSF shifts on serial head CTs. / Chen, Yasheng; Dhar, Rajat; Heitsch, Laura; Ford, Andria; Fernandez-Cadenas, Israel; Carrera, Caty; Montaner, Joan; Lin, Weili; Shen, Dinggang; An, Hongyu; Lee, Jin Moo.

In: NeuroImage: Clinical, Vol. 12, 2016, p. 673-680.

Research output: Contribution to journalArticle

Chen, Yasheng ; Dhar, Rajat ; Heitsch, Laura ; Ford, Andria ; Fernandez-Cadenas, Israel ; Carrera, Caty ; Montaner, Joan ; Lin, Weili ; Shen, Dinggang ; An, Hongyu ; Lee, Jin Moo. / Automated quantification of cerebral edema following hemispheric infarction : Application of a machine-learning algorithm to evaluate CSF shifts on serial head CTs. In: NeuroImage: Clinical. 2016 ; Vol. 12. pp. 673-680.
@article{ac56b7e2ad4f43deb3716566ab65753a,
title = "Automated quantification of cerebral edema following hemispheric infarction: Application of a machine-learning algorithm to evaluate CSF shifts on serial head CTs",
abstract = "Although cerebral edema is a major cause of death and deterioration following hemispheric stroke, there remains no validated biomarker that captures the full spectrum of this critical complication. We recently demonstrated that reduction in intracranial cerebrospinal fluid (CSF) volume (∆ CSF) on serial computed tomography (CT) scans provides an accurate measure of cerebral edema severity, which may aid in early triaging of stroke patients for craniectomy. However, application of such a volumetric approach would be too cumbersome to perform manually on serial scans in a real-world setting. We developed and validated an automated technique for CSF segmentation via integration of random forest (RF) based machine learning with geodesic active contour (GAC) segmentation. The proposed RF + GAC approach was compared to conventional Hounsfield Unit (HU) thresholding and RF segmentation methods using Dice similarity coefficient (DSC) and the correlation of volumetric measurements, with manual delineation serving as the ground truth. CSF spaces were outlined on scans performed at baseline (< 6 h after stroke onset) and early follow-up (FU) (closest to 24 h) in 38 acute ischemic stroke patients. RF performed significantly better than optimized HU thresholding (p < 10− 4 in baseline and p < 10− 5 in FU) and RF + GAC performed significantly better than RF (p < 10− 3 in baseline and p < 10− 5 in FU). Pearson correlation coefficients between the automatically detected ∆ CSF and the ground truth were r = 0.178 (p = 0.285), r = 0.876 (p < 10− 6) and r = 0.879 (p < 10− 6) for thresholding, RF and RF + GAC, respectively, with a slope closer to the line of identity in RF + GAC. When we applied the algorithm trained from images of one stroke center to segment CTs from another center, similar findings held. In conclusion, we have developed and validated an accurate automated approach to segment CSF and calculate its shifts on serial CT scans. This algorithm will allow us to efficiently and accurately measure the evolution of cerebral edema in future studies including large multi-site patient populations.",
keywords = "Active contour, Cerebral edema, CSF segmentation, Ischemic stroke CT, Mass effect, Random forest",
author = "Yasheng Chen and Rajat Dhar and Laura Heitsch and Andria Ford and Israel Fernandez-Cadenas and Caty Carrera and Joan Montaner and Weili Lin and Dinggang Shen and Hongyu An and Lee, {Jin Moo}",
year = "2016",
doi = "10.1016/j.nicl.2016.09.018",
language = "English",
volume = "12",
pages = "673--680",
journal = "NeuroImage: Clinical",
issn = "2213-1582",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Automated quantification of cerebral edema following hemispheric infarction

T2 - Application of a machine-learning algorithm to evaluate CSF shifts on serial head CTs

AU - Chen, Yasheng

AU - Dhar, Rajat

AU - Heitsch, Laura

AU - Ford, Andria

AU - Fernandez-Cadenas, Israel

AU - Carrera, Caty

AU - Montaner, Joan

AU - Lin, Weili

AU - Shen, Dinggang

AU - An, Hongyu

AU - Lee, Jin Moo

PY - 2016

Y1 - 2016

N2 - Although cerebral edema is a major cause of death and deterioration following hemispheric stroke, there remains no validated biomarker that captures the full spectrum of this critical complication. We recently demonstrated that reduction in intracranial cerebrospinal fluid (CSF) volume (∆ CSF) on serial computed tomography (CT) scans provides an accurate measure of cerebral edema severity, which may aid in early triaging of stroke patients for craniectomy. However, application of such a volumetric approach would be too cumbersome to perform manually on serial scans in a real-world setting. We developed and validated an automated technique for CSF segmentation via integration of random forest (RF) based machine learning with geodesic active contour (GAC) segmentation. The proposed RF + GAC approach was compared to conventional Hounsfield Unit (HU) thresholding and RF segmentation methods using Dice similarity coefficient (DSC) and the correlation of volumetric measurements, with manual delineation serving as the ground truth. CSF spaces were outlined on scans performed at baseline (< 6 h after stroke onset) and early follow-up (FU) (closest to 24 h) in 38 acute ischemic stroke patients. RF performed significantly better than optimized HU thresholding (p < 10− 4 in baseline and p < 10− 5 in FU) and RF + GAC performed significantly better than RF (p < 10− 3 in baseline and p < 10− 5 in FU). Pearson correlation coefficients between the automatically detected ∆ CSF and the ground truth were r = 0.178 (p = 0.285), r = 0.876 (p < 10− 6) and r = 0.879 (p < 10− 6) for thresholding, RF and RF + GAC, respectively, with a slope closer to the line of identity in RF + GAC. When we applied the algorithm trained from images of one stroke center to segment CTs from another center, similar findings held. In conclusion, we have developed and validated an accurate automated approach to segment CSF and calculate its shifts on serial CT scans. This algorithm will allow us to efficiently and accurately measure the evolution of cerebral edema in future studies including large multi-site patient populations.

AB - Although cerebral edema is a major cause of death and deterioration following hemispheric stroke, there remains no validated biomarker that captures the full spectrum of this critical complication. We recently demonstrated that reduction in intracranial cerebrospinal fluid (CSF) volume (∆ CSF) on serial computed tomography (CT) scans provides an accurate measure of cerebral edema severity, which may aid in early triaging of stroke patients for craniectomy. However, application of such a volumetric approach would be too cumbersome to perform manually on serial scans in a real-world setting. We developed and validated an automated technique for CSF segmentation via integration of random forest (RF) based machine learning with geodesic active contour (GAC) segmentation. The proposed RF + GAC approach was compared to conventional Hounsfield Unit (HU) thresholding and RF segmentation methods using Dice similarity coefficient (DSC) and the correlation of volumetric measurements, with manual delineation serving as the ground truth. CSF spaces were outlined on scans performed at baseline (< 6 h after stroke onset) and early follow-up (FU) (closest to 24 h) in 38 acute ischemic stroke patients. RF performed significantly better than optimized HU thresholding (p < 10− 4 in baseline and p < 10− 5 in FU) and RF + GAC performed significantly better than RF (p < 10− 3 in baseline and p < 10− 5 in FU). Pearson correlation coefficients between the automatically detected ∆ CSF and the ground truth were r = 0.178 (p = 0.285), r = 0.876 (p < 10− 6) and r = 0.879 (p < 10− 6) for thresholding, RF and RF + GAC, respectively, with a slope closer to the line of identity in RF + GAC. When we applied the algorithm trained from images of one stroke center to segment CTs from another center, similar findings held. In conclusion, we have developed and validated an accurate automated approach to segment CSF and calculate its shifts on serial CT scans. This algorithm will allow us to efficiently and accurately measure the evolution of cerebral edema in future studies including large multi-site patient populations.

KW - Active contour

KW - Cerebral edema

KW - CSF segmentation

KW - Ischemic stroke CT

KW - Mass effect

KW - Random forest

UR - http://www.scopus.com/inward/record.url?scp=84991463718&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84991463718&partnerID=8YFLogxK

U2 - 10.1016/j.nicl.2016.09.018

DO - 10.1016/j.nicl.2016.09.018

M3 - Article

VL - 12

SP - 673

EP - 680

JO - NeuroImage: Clinical

JF - NeuroImage: Clinical

SN - 2213-1582

ER -