Automated and manual hippocampal segmentation techniques

Comparison of results, reproducibility and clinical applicability

Sona Hurtz, Nicole Chow, Amity E. Watson, Johanne H. Somme, Naira Goukasian, Kristy S. Hwang, John Morra, David Elashoff, Sujuan Gao, Ronald C. Petersen, Paul S. Aisen, Paul M. Thompson, Liana G. Apostolova

Research output: Contribution to journalArticle

Abstract

Background: Imaging techniques used to measure hippocampal atrophy are key to understanding the clinical progression of Alzheimer's disease (AD). Various semi-automated hippocampal segmentation techniques are available and require human expert input to learn how to accurately segment new data. Our goal was to compare 1) the performance of our automated hippocampal segmentation technique relative to manual segmentations, and 2) the performance of our automated technique when provided with a training set from two different raters. We also explored the ability of hippocampal volumes obtained using manual and automated hippocampal segmentations to predict conversion from MCI to AD. Methods: We analyzed 161 1.5 T T1-weighted brain magnetic resonance images (MRI) from the ADCS Donepezil/Vitamin E clinical study. All subjects carried a diagnosis of mild cognitive impairment (MCI). Three different segmentation outputs (one produced by manual tracing and two produced by a semi-automated algorithm trained with training sets developed by two raters) were compared using single measure intraclass correlation statistics (smICC). The radial distance method was used to assess each segmentation technique's ability to detect hippocampal atrophy in 3D. We then compared how well each segmentation method detected baseline hippocampal differences between MCI subjects who remained stable (MCInc) and those who converted to AD (MCIc) during the trial. Our statistical maps were corrected for multiple comparisons using permutation-based statistics with a threshold of p < .01. Results: Our smICC analyses showed significant agreement between the manual and automated hippocampal segmentations from rater 1 [right smICC = 0.78 (95%CI 0.72–0.84); left smICC = 0.79 (95%CI 0.72–0.85)], the manual segmentations from rater 1 versus the automated segmentations from rater 2 [right smICC = 0.78 (95%CI 0.7–0.84); left smICC = 0.78 (95%CI 0.71–0.84)], and the automated segmentations of rater 1 versus rater 2 [right smICC = 0.97 (95%CI 0.96–0.98); left smICC = 0.97 (95%CI 0.96–0.98)]. All three segmentation methods detected significant CA1 and subicular atrophy in MCIc compared to MCInc at baseline (manual: right pcorrected = 0.0112, left pcorrected = 0.0006; automated rater 1: right pcorrected = 0.0318, left pcorrected = 0.0302; automated rater 2: right pcorrected = 0.0029, left pcorrected = 0.0166). Conclusions: The hippocampal volumes obtained with a fast semi-automated segmentation method were highly comparable to the ones obtained with the labor-intensive manual segmentation method. The AdaBoost automated hippocampal segmentation technique is highly reliable allowing the efficient analysis of large data sets.

Original languageEnglish (US)
Article number101574
JournalNeuroImage: Clinical
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Reproducibility of Results
Atrophy
Alzheimer Disease
Aptitude
Vitamin E
Magnetic Resonance Spectroscopy
Brain

Keywords

  • Alzheimer's disease
  • Automated segmentation
  • Magnetic resonance imaging
  • Manual segmentation
  • Mild cognitive impairment
  • Radial distance mapping

ASJC Scopus subject areas

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

Cite this

Hurtz, S., Chow, N., Watson, A. E., Somme, J. H., Goukasian, N., Hwang, K. S., ... Apostolova, L. G. (Accepted/In press). Automated and manual hippocampal segmentation techniques: Comparison of results, reproducibility and clinical applicability. NeuroImage: Clinical, [101574]. https://doi.org/10.1016/j.nicl.2018.10.012

Automated and manual hippocampal segmentation techniques : Comparison of results, reproducibility and clinical applicability. / Hurtz, Sona; Chow, Nicole; Watson, Amity E.; Somme, Johanne H.; Goukasian, Naira; Hwang, Kristy S.; Morra, John; Elashoff, David; Gao, Sujuan; Petersen, Ronald C.; Aisen, Paul S.; Thompson, Paul M.; Apostolova, Liana G.

In: NeuroImage: Clinical, 01.01.2018.

Research output: Contribution to journalArticle

Hurtz, S, Chow, N, Watson, AE, Somme, JH, Goukasian, N, Hwang, KS, Morra, J, Elashoff, D, Gao, S, Petersen, RC, Aisen, PS, Thompson, PM & Apostolova, LG 2018, 'Automated and manual hippocampal segmentation techniques: Comparison of results, reproducibility and clinical applicability', NeuroImage: Clinical. https://doi.org/10.1016/j.nicl.2018.10.012
Hurtz, Sona ; Chow, Nicole ; Watson, Amity E. ; Somme, Johanne H. ; Goukasian, Naira ; Hwang, Kristy S. ; Morra, John ; Elashoff, David ; Gao, Sujuan ; Petersen, Ronald C. ; Aisen, Paul S. ; Thompson, Paul M. ; Apostolova, Liana G. / Automated and manual hippocampal segmentation techniques : Comparison of results, reproducibility and clinical applicability. In: NeuroImage: Clinical. 2018.
@article{4f5b2fc1adc049ee878640622f4f5d8c,
title = "Automated and manual hippocampal segmentation techniques: Comparison of results, reproducibility and clinical applicability",
abstract = "Background: Imaging techniques used to measure hippocampal atrophy are key to understanding the clinical progression of Alzheimer's disease (AD). Various semi-automated hippocampal segmentation techniques are available and require human expert input to learn how to accurately segment new data. Our goal was to compare 1) the performance of our automated hippocampal segmentation technique relative to manual segmentations, and 2) the performance of our automated technique when provided with a training set from two different raters. We also explored the ability of hippocampal volumes obtained using manual and automated hippocampal segmentations to predict conversion from MCI to AD. Methods: We analyzed 161 1.5 T T1-weighted brain magnetic resonance images (MRI) from the ADCS Donepezil/Vitamin E clinical study. All subjects carried a diagnosis of mild cognitive impairment (MCI). Three different segmentation outputs (one produced by manual tracing and two produced by a semi-automated algorithm trained with training sets developed by two raters) were compared using single measure intraclass correlation statistics (smICC). The radial distance method was used to assess each segmentation technique's ability to detect hippocampal atrophy in 3D. We then compared how well each segmentation method detected baseline hippocampal differences between MCI subjects who remained stable (MCInc) and those who converted to AD (MCIc) during the trial. Our statistical maps were corrected for multiple comparisons using permutation-based statistics with a threshold of p < .01. Results: Our smICC analyses showed significant agreement between the manual and automated hippocampal segmentations from rater 1 [right smICC = 0.78 (95{\%}CI 0.72–0.84); left smICC = 0.79 (95{\%}CI 0.72–0.85)], the manual segmentations from rater 1 versus the automated segmentations from rater 2 [right smICC = 0.78 (95{\%}CI 0.7–0.84); left smICC = 0.78 (95{\%}CI 0.71–0.84)], and the automated segmentations of rater 1 versus rater 2 [right smICC = 0.97 (95{\%}CI 0.96–0.98); left smICC = 0.97 (95{\%}CI 0.96–0.98)]. All three segmentation methods detected significant CA1 and subicular atrophy in MCIc compared to MCInc at baseline (manual: right pcorrected = 0.0112, left pcorrected = 0.0006; automated rater 1: right pcorrected = 0.0318, left pcorrected = 0.0302; automated rater 2: right pcorrected = 0.0029, left pcorrected = 0.0166). Conclusions: The hippocampal volumes obtained with a fast semi-automated segmentation method were highly comparable to the ones obtained with the labor-intensive manual segmentation method. The AdaBoost automated hippocampal segmentation technique is highly reliable allowing the efficient analysis of large data sets.",
keywords = "Alzheimer's disease, Automated segmentation, Magnetic resonance imaging, Manual segmentation, Mild cognitive impairment, Radial distance mapping",
author = "Sona Hurtz and Nicole Chow and Watson, {Amity E.} and Somme, {Johanne H.} and Naira Goukasian and Hwang, {Kristy S.} and John Morra and David Elashoff and Sujuan Gao and Petersen, {Ronald C.} and Aisen, {Paul S.} and Thompson, {Paul M.} and Apostolova, {Liana G.}",
year = "2018",
month = "1",
day = "1",
doi = "10.1016/j.nicl.2018.10.012",
language = "English (US)",
journal = "NeuroImage: Clinical",
issn = "2213-1582",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Automated and manual hippocampal segmentation techniques

T2 - Comparison of results, reproducibility and clinical applicability

AU - Hurtz, Sona

AU - Chow, Nicole

AU - Watson, Amity E.

AU - Somme, Johanne H.

AU - Goukasian, Naira

AU - Hwang, Kristy S.

AU - Morra, John

AU - Elashoff, David

AU - Gao, Sujuan

AU - Petersen, Ronald C.

AU - Aisen, Paul S.

AU - Thompson, Paul M.

AU - Apostolova, Liana G.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Background: Imaging techniques used to measure hippocampal atrophy are key to understanding the clinical progression of Alzheimer's disease (AD). Various semi-automated hippocampal segmentation techniques are available and require human expert input to learn how to accurately segment new data. Our goal was to compare 1) the performance of our automated hippocampal segmentation technique relative to manual segmentations, and 2) the performance of our automated technique when provided with a training set from two different raters. We also explored the ability of hippocampal volumes obtained using manual and automated hippocampal segmentations to predict conversion from MCI to AD. Methods: We analyzed 161 1.5 T T1-weighted brain magnetic resonance images (MRI) from the ADCS Donepezil/Vitamin E clinical study. All subjects carried a diagnosis of mild cognitive impairment (MCI). Three different segmentation outputs (one produced by manual tracing and two produced by a semi-automated algorithm trained with training sets developed by two raters) were compared using single measure intraclass correlation statistics (smICC). The radial distance method was used to assess each segmentation technique's ability to detect hippocampal atrophy in 3D. We then compared how well each segmentation method detected baseline hippocampal differences between MCI subjects who remained stable (MCInc) and those who converted to AD (MCIc) during the trial. Our statistical maps were corrected for multiple comparisons using permutation-based statistics with a threshold of p < .01. Results: Our smICC analyses showed significant agreement between the manual and automated hippocampal segmentations from rater 1 [right smICC = 0.78 (95%CI 0.72–0.84); left smICC = 0.79 (95%CI 0.72–0.85)], the manual segmentations from rater 1 versus the automated segmentations from rater 2 [right smICC = 0.78 (95%CI 0.7–0.84); left smICC = 0.78 (95%CI 0.71–0.84)], and the automated segmentations of rater 1 versus rater 2 [right smICC = 0.97 (95%CI 0.96–0.98); left smICC = 0.97 (95%CI 0.96–0.98)]. All three segmentation methods detected significant CA1 and subicular atrophy in MCIc compared to MCInc at baseline (manual: right pcorrected = 0.0112, left pcorrected = 0.0006; automated rater 1: right pcorrected = 0.0318, left pcorrected = 0.0302; automated rater 2: right pcorrected = 0.0029, left pcorrected = 0.0166). Conclusions: The hippocampal volumes obtained with a fast semi-automated segmentation method were highly comparable to the ones obtained with the labor-intensive manual segmentation method. The AdaBoost automated hippocampal segmentation technique is highly reliable allowing the efficient analysis of large data sets.

AB - Background: Imaging techniques used to measure hippocampal atrophy are key to understanding the clinical progression of Alzheimer's disease (AD). Various semi-automated hippocampal segmentation techniques are available and require human expert input to learn how to accurately segment new data. Our goal was to compare 1) the performance of our automated hippocampal segmentation technique relative to manual segmentations, and 2) the performance of our automated technique when provided with a training set from two different raters. We also explored the ability of hippocampal volumes obtained using manual and automated hippocampal segmentations to predict conversion from MCI to AD. Methods: We analyzed 161 1.5 T T1-weighted brain magnetic resonance images (MRI) from the ADCS Donepezil/Vitamin E clinical study. All subjects carried a diagnosis of mild cognitive impairment (MCI). Three different segmentation outputs (one produced by manual tracing and two produced by a semi-automated algorithm trained with training sets developed by two raters) were compared using single measure intraclass correlation statistics (smICC). The radial distance method was used to assess each segmentation technique's ability to detect hippocampal atrophy in 3D. We then compared how well each segmentation method detected baseline hippocampal differences between MCI subjects who remained stable (MCInc) and those who converted to AD (MCIc) during the trial. Our statistical maps were corrected for multiple comparisons using permutation-based statistics with a threshold of p < .01. Results: Our smICC analyses showed significant agreement between the manual and automated hippocampal segmentations from rater 1 [right smICC = 0.78 (95%CI 0.72–0.84); left smICC = 0.79 (95%CI 0.72–0.85)], the manual segmentations from rater 1 versus the automated segmentations from rater 2 [right smICC = 0.78 (95%CI 0.7–0.84); left smICC = 0.78 (95%CI 0.71–0.84)], and the automated segmentations of rater 1 versus rater 2 [right smICC = 0.97 (95%CI 0.96–0.98); left smICC = 0.97 (95%CI 0.96–0.98)]. All three segmentation methods detected significant CA1 and subicular atrophy in MCIc compared to MCInc at baseline (manual: right pcorrected = 0.0112, left pcorrected = 0.0006; automated rater 1: right pcorrected = 0.0318, left pcorrected = 0.0302; automated rater 2: right pcorrected = 0.0029, left pcorrected = 0.0166). Conclusions: The hippocampal volumes obtained with a fast semi-automated segmentation method were highly comparable to the ones obtained with the labor-intensive manual segmentation method. The AdaBoost automated hippocampal segmentation technique is highly reliable allowing the efficient analysis of large data sets.

KW - Alzheimer's disease

KW - Automated segmentation

KW - Magnetic resonance imaging

KW - Manual segmentation

KW - Mild cognitive impairment

KW - Radial distance mapping

UR - http://www.scopus.com/inward/record.url?scp=85058188361&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058188361&partnerID=8YFLogxK

U2 - 10.1016/j.nicl.2018.10.012

DO - 10.1016/j.nicl.2018.10.012

M3 - Article

JO - NeuroImage: Clinical

JF - NeuroImage: Clinical

SN - 2213-1582

M1 - 101574

ER -