Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI

Ying Wu, Simon K. Warfield, I. Leng Tan, William M. Wells III, Dominik S. Meier, Ronald A. van Schijndel, Frederik Barkhof, Charles R. G. Guttmann

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Purpose.: To automatically segment multiple sclerosis (MS) lesions into three subtypes (i.e., enhancing lesions, T1 "black holes", T2 hyperintense lesions). Materials and methods.: Proton density-, T2- and contrast-enhanced T1-weighted brain images of 12 MR scans were pre-processed through intracranial cavity (IC) extraction, inhomogeneity correction and intensity normalization. Intensity-based statistical k-nearest neighbor (k-NN) classification was combined with template-driven segmentation and partial volume artifact correction (TDS+) for segmentation of MS lesions subtypes and brain tissue compartments. Operator-supervised tissue sampling and parameter calibration were performed on 2 randomly selected scans and were applied automatically to the remaining 10 scans. Results from this three-channel TDS+ (3ch-TDS+) were compared to those from a previously validated two-channel TDS+ (2ch-TDS+) method. The results of both the 3ch-TDS+ and 2ch-TDS+ were also compared to manual segmentation performed by experts. Results.: Intra-class correlation coefficients (ICC) of 3ch-TDS+ for all three subtypes of lesions were higher (ICC between 0.95 and 0.96) than that of 2ch-TDS+ for T2 lesions (ICC = 0.82). The 3ch-TDS+ also identified the three lesion subtypes with high specificity (98.7-99.9%) and accuracy (98.5-99.9%). Sensitivity of 3ch-TDS+ for T2 lesions was 16% higher than with 2ch-TDS+. Enhancing lesions were segmented with the best sensitivity (81.9%). "Black holes" were segmented with the least sensitivity (62.3%). Conclusion.: 3ch-TDS+ is a promising method for automated segmentation of MS lesion subtypes. © 2006 Elsevier Inc. All rights reserved.
Original languageEnglish
Pages (from-to)1205-1215
JournalNeuroImage
Volume32
Issue number3
DOIs
Publication statusPublished - 2006

Cite this

Wu, Y., Warfield, S. K., Tan, I. L., Wells III, W. M., Meier, D. S., van Schijndel, R. A., ... Guttmann, C. R. G. (2006). Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI. NeuroImage, 32(3), 1205-1215. https://doi.org/10.1016/j.neuroimage.2006.04.211
Wu, Ying ; Warfield, Simon K. ; Tan, I. Leng ; Wells III, William M. ; Meier, Dominik S. ; van Schijndel, Ronald A. ; Barkhof, Frederik ; Guttmann, Charles R. G. / Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI. In: NeuroImage. 2006 ; Vol. 32, No. 3. pp. 1205-1215.
@article{6f73c8488f1a45898b1b1cfd3c904470,
title = "Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI",
abstract = "Purpose.: To automatically segment multiple sclerosis (MS) lesions into three subtypes (i.e., enhancing lesions, T1 {"}black holes{"}, T2 hyperintense lesions). Materials and methods.: Proton density-, T2- and contrast-enhanced T1-weighted brain images of 12 MR scans were pre-processed through intracranial cavity (IC) extraction, inhomogeneity correction and intensity normalization. Intensity-based statistical k-nearest neighbor (k-NN) classification was combined with template-driven segmentation and partial volume artifact correction (TDS+) for segmentation of MS lesions subtypes and brain tissue compartments. Operator-supervised tissue sampling and parameter calibration were performed on 2 randomly selected scans and were applied automatically to the remaining 10 scans. Results from this three-channel TDS+ (3ch-TDS+) were compared to those from a previously validated two-channel TDS+ (2ch-TDS+) method. The results of both the 3ch-TDS+ and 2ch-TDS+ were also compared to manual segmentation performed by experts. Results.: Intra-class correlation coefficients (ICC) of 3ch-TDS+ for all three subtypes of lesions were higher (ICC between 0.95 and 0.96) than that of 2ch-TDS+ for T2 lesions (ICC = 0.82). The 3ch-TDS+ also identified the three lesion subtypes with high specificity (98.7-99.9{\%}) and accuracy (98.5-99.9{\%}). Sensitivity of 3ch-TDS+ for T2 lesions was 16{\%} higher than with 2ch-TDS+. Enhancing lesions were segmented with the best sensitivity (81.9{\%}). {"}Black holes{"} were segmented with the least sensitivity (62.3{\%}). Conclusion.: 3ch-TDS+ is a promising method for automated segmentation of MS lesion subtypes. {\circledC} 2006 Elsevier Inc. All rights reserved.",
author = "Ying Wu and Warfield, {Simon K.} and Tan, {I. Leng} and {Wells III}, {William M.} and Meier, {Dominik S.} and {van Schijndel}, {Ronald A.} and Frederik Barkhof and Guttmann, {Charles R. G.}",
year = "2006",
doi = "10.1016/j.neuroimage.2006.04.211",
language = "English",
volume = "32",
pages = "1205--1215",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",
number = "3",

}

Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI. / Wu, Ying; Warfield, Simon K.; Tan, I. Leng; Wells III, William M.; Meier, Dominik S.; van Schijndel, Ronald A.; Barkhof, Frederik; Guttmann, Charles R. G.

In: NeuroImage, Vol. 32, No. 3, 2006, p. 1205-1215.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI

AU - Wu, Ying

AU - Warfield, Simon K.

AU - Tan, I. Leng

AU - Wells III, William M.

AU - Meier, Dominik S.

AU - van Schijndel, Ronald A.

AU - Barkhof, Frederik

AU - Guttmann, Charles R. G.

PY - 2006

Y1 - 2006

N2 - Purpose.: To automatically segment multiple sclerosis (MS) lesions into three subtypes (i.e., enhancing lesions, T1 "black holes", T2 hyperintense lesions). Materials and methods.: Proton density-, T2- and contrast-enhanced T1-weighted brain images of 12 MR scans were pre-processed through intracranial cavity (IC) extraction, inhomogeneity correction and intensity normalization. Intensity-based statistical k-nearest neighbor (k-NN) classification was combined with template-driven segmentation and partial volume artifact correction (TDS+) for segmentation of MS lesions subtypes and brain tissue compartments. Operator-supervised tissue sampling and parameter calibration were performed on 2 randomly selected scans and were applied automatically to the remaining 10 scans. Results from this three-channel TDS+ (3ch-TDS+) were compared to those from a previously validated two-channel TDS+ (2ch-TDS+) method. The results of both the 3ch-TDS+ and 2ch-TDS+ were also compared to manual segmentation performed by experts. Results.: Intra-class correlation coefficients (ICC) of 3ch-TDS+ for all three subtypes of lesions were higher (ICC between 0.95 and 0.96) than that of 2ch-TDS+ for T2 lesions (ICC = 0.82). The 3ch-TDS+ also identified the three lesion subtypes with high specificity (98.7-99.9%) and accuracy (98.5-99.9%). Sensitivity of 3ch-TDS+ for T2 lesions was 16% higher than with 2ch-TDS+. Enhancing lesions were segmented with the best sensitivity (81.9%). "Black holes" were segmented with the least sensitivity (62.3%). Conclusion.: 3ch-TDS+ is a promising method for automated segmentation of MS lesion subtypes. © 2006 Elsevier Inc. All rights reserved.

AB - Purpose.: To automatically segment multiple sclerosis (MS) lesions into three subtypes (i.e., enhancing lesions, T1 "black holes", T2 hyperintense lesions). Materials and methods.: Proton density-, T2- and contrast-enhanced T1-weighted brain images of 12 MR scans were pre-processed through intracranial cavity (IC) extraction, inhomogeneity correction and intensity normalization. Intensity-based statistical k-nearest neighbor (k-NN) classification was combined with template-driven segmentation and partial volume artifact correction (TDS+) for segmentation of MS lesions subtypes and brain tissue compartments. Operator-supervised tissue sampling and parameter calibration were performed on 2 randomly selected scans and were applied automatically to the remaining 10 scans. Results from this three-channel TDS+ (3ch-TDS+) were compared to those from a previously validated two-channel TDS+ (2ch-TDS+) method. The results of both the 3ch-TDS+ and 2ch-TDS+ were also compared to manual segmentation performed by experts. Results.: Intra-class correlation coefficients (ICC) of 3ch-TDS+ for all three subtypes of lesions were higher (ICC between 0.95 and 0.96) than that of 2ch-TDS+ for T2 lesions (ICC = 0.82). The 3ch-TDS+ also identified the three lesion subtypes with high specificity (98.7-99.9%) and accuracy (98.5-99.9%). Sensitivity of 3ch-TDS+ for T2 lesions was 16% higher than with 2ch-TDS+. Enhancing lesions were segmented with the best sensitivity (81.9%). "Black holes" were segmented with the least sensitivity (62.3%). Conclusion.: 3ch-TDS+ is a promising method for automated segmentation of MS lesion subtypes. © 2006 Elsevier Inc. All rights reserved.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=33748062327&origin=inward

UR - https://www.ncbi.nlm.nih.gov/pubmed/16797188

U2 - 10.1016/j.neuroimage.2006.04.211

DO - 10.1016/j.neuroimage.2006.04.211

M3 - Article

VL - 32

SP - 1205

EP - 1215

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

IS - 3

ER -