Similarity clustering-based atlas selection for pelvic CT image segmentation

Angel Kennedy, Jason Dowling, Peter B. Greer, Lois Holloway, Michael G. Jameson, Dale Roach, Soumya Ghose, David Rivest-Henault, Marco Marcello, Martin A. Ebert

Research output: Contribution to journalArticle

Abstract

Purpose To demonstrate selection of a small representative subset of images from a pool of images comprising a potential atlas (PA) pelvic CT set to be used for autosegmentation of a separate target image set. The aim is to balance the need for the atlas set to represent anatomical diversity with the need to minimize resources required to create a high quality atlas set (such as multiobserver delineation), while retaining access to additional information available for the PA image set. Methods Preprocessing was performed for image standardization, followed by image registration. Clustering was used to select the subset that provided the best coverage of a target dataset as measured by postregistration image intensity similarities. Tests for clustering robustness were performed including repeated clustering runs using different starting seeds and clustering repeatedly using 90% of the target dataset chosen randomly. Comparisons of coverage of a target set (comprising 711 pelvic CT images) were made for atlas sets of five images (chosen from a PA set of 39 pelvic CT and MR images) (a) at random (averaged over 50 random atlas selections), (b) based solely on image similarities within the PA set (representing prospective atlas development), (c) based on similarities within the PA set and between the PA and target dataset (representing retrospective atlas development). Comparisons were also made to coverage provided by the entire PA set of 39 images. Results Exemplar selection was highly robust with exemplar selection results being unaffected by choice of starting seed with very occasional change to one of the exemplar choices when the target set was reduced. Coverage of the target set, as measured by best normalized cross-correlation similarity of target images to any exemplar image, provided by five well-selected atlas images (mean = 0.6497) was more similar to coverage provided by the entire PA set (mean = 0.6658) than randomly chosen atlas subsets (mean = 0.5977). This was true both of the mean values and the shape of the distributions. Retrospective selection of atlases (mean = 0.6497) provided a very small improvement over prospective atlas selection (mean = 0.6431). All differences were significant (P <1.0E-10). Conclusions Selection of a small representative image set from one dataset can be utilized to develop an atlas set for either retrospective or prospective autosegmentation of a different target dataset. The coverage provided by such a judiciously selected subset has the potential to facilitate propagation of numerous retrospectively defined structures, utilizing additional information available with multimodal imaging in the atlas set, without the need to create large atlas image sets.

Original languageEnglish
Pages (from-to)2243-2250
Number of pages8
JournalMedical Physics
Volume46
Issue number5
DOIs
Publication statusPublished - May 2019

Cite this

Kennedy, A., Dowling, J., Greer, P. B., Holloway, L., Jameson, M. G., Roach, D., ... Ebert, M. A. (2019). Similarity clustering-based atlas selection for pelvic CT image segmentation. Medical Physics, 46(5), 2243-2250. https://doi.org/10.1002/mp.13494
Kennedy, Angel ; Dowling, Jason ; Greer, Peter B. ; Holloway, Lois ; Jameson, Michael G. ; Roach, Dale ; Ghose, Soumya ; Rivest-Henault, David ; Marcello, Marco ; Ebert, Martin A. / Similarity clustering-based atlas selection for pelvic CT image segmentation. In: Medical Physics. 2019 ; Vol. 46, No. 5. pp. 2243-2250.
@article{ae6580bea3c147daa820f4b66cdd714d,
title = "Similarity clustering-based atlas selection for pelvic CT image segmentation",
abstract = "Purpose To demonstrate selection of a small representative subset of images from a pool of images comprising a potential atlas (PA) pelvic CT set to be used for autosegmentation of a separate target image set. The aim is to balance the need for the atlas set to represent anatomical diversity with the need to minimize resources required to create a high quality atlas set (such as multiobserver delineation), while retaining access to additional information available for the PA image set. Methods Preprocessing was performed for image standardization, followed by image registration. Clustering was used to select the subset that provided the best coverage of a target dataset as measured by postregistration image intensity similarities. Tests for clustering robustness were performed including repeated clustering runs using different starting seeds and clustering repeatedly using 90{\%} of the target dataset chosen randomly. Comparisons of coverage of a target set (comprising 711 pelvic CT images) were made for atlas sets of five images (chosen from a PA set of 39 pelvic CT and MR images) (a) at random (averaged over 50 random atlas selections), (b) based solely on image similarities within the PA set (representing prospective atlas development), (c) based on similarities within the PA set and between the PA and target dataset (representing retrospective atlas development). Comparisons were also made to coverage provided by the entire PA set of 39 images. Results Exemplar selection was highly robust with exemplar selection results being unaffected by choice of starting seed with very occasional change to one of the exemplar choices when the target set was reduced. Coverage of the target set, as measured by best normalized cross-correlation similarity of target images to any exemplar image, provided by five well-selected atlas images (mean = 0.6497) was more similar to coverage provided by the entire PA set (mean = 0.6658) than randomly chosen atlas subsets (mean = 0.5977). This was true both of the mean values and the shape of the distributions. Retrospective selection of atlases (mean = 0.6497) provided a very small improvement over prospective atlas selection (mean = 0.6431). All differences were significant (P <1.0E-10). Conclusions Selection of a small representative image set from one dataset can be utilized to develop an atlas set for either retrospective or prospective autosegmentation of a different target dataset. The coverage provided by such a judiciously selected subset has the potential to facilitate propagation of numerous retrospectively defined structures, utilizing additional information available with multimodal imaging in the atlas set, without the need to create large atlas image sets.",
keywords = "autosegmentation, clustering, image-atlas, image registration, REGISTRATION, THERAPY",
author = "Angel Kennedy and Jason Dowling and Greer, {Peter B.} and Lois Holloway and Jameson, {Michael G.} and Dale Roach and Soumya Ghose and David Rivest-Henault and Marco Marcello and Ebert, {Martin A.}",
year = "2019",
month = "5",
doi = "10.1002/mp.13494",
language = "English",
volume = "46",
pages = "2243--2250",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "American Institute of Physics",
number = "5",

}

Kennedy, A, Dowling, J, Greer, PB, Holloway, L, Jameson, MG, Roach, D, Ghose, S, Rivest-Henault, D, Marcello, M & Ebert, MA 2019, 'Similarity clustering-based atlas selection for pelvic CT image segmentation' Medical Physics, vol. 46, no. 5, pp. 2243-2250. https://doi.org/10.1002/mp.13494

Similarity clustering-based atlas selection for pelvic CT image segmentation. / Kennedy, Angel; Dowling, Jason; Greer, Peter B.; Holloway, Lois; Jameson, Michael G.; Roach, Dale; Ghose, Soumya; Rivest-Henault, David; Marcello, Marco; Ebert, Martin A.

In: Medical Physics, Vol. 46, No. 5, 05.2019, p. 2243-2250.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Similarity clustering-based atlas selection for pelvic CT image segmentation

AU - Kennedy, Angel

AU - Dowling, Jason

AU - Greer, Peter B.

AU - Holloway, Lois

AU - Jameson, Michael G.

AU - Roach, Dale

AU - Ghose, Soumya

AU - Rivest-Henault, David

AU - Marcello, Marco

AU - Ebert, Martin A.

PY - 2019/5

Y1 - 2019/5

N2 - Purpose To demonstrate selection of a small representative subset of images from a pool of images comprising a potential atlas (PA) pelvic CT set to be used for autosegmentation of a separate target image set. The aim is to balance the need for the atlas set to represent anatomical diversity with the need to minimize resources required to create a high quality atlas set (such as multiobserver delineation), while retaining access to additional information available for the PA image set. Methods Preprocessing was performed for image standardization, followed by image registration. Clustering was used to select the subset that provided the best coverage of a target dataset as measured by postregistration image intensity similarities. Tests for clustering robustness were performed including repeated clustering runs using different starting seeds and clustering repeatedly using 90% of the target dataset chosen randomly. Comparisons of coverage of a target set (comprising 711 pelvic CT images) were made for atlas sets of five images (chosen from a PA set of 39 pelvic CT and MR images) (a) at random (averaged over 50 random atlas selections), (b) based solely on image similarities within the PA set (representing prospective atlas development), (c) based on similarities within the PA set and between the PA and target dataset (representing retrospective atlas development). Comparisons were also made to coverage provided by the entire PA set of 39 images. Results Exemplar selection was highly robust with exemplar selection results being unaffected by choice of starting seed with very occasional change to one of the exemplar choices when the target set was reduced. Coverage of the target set, as measured by best normalized cross-correlation similarity of target images to any exemplar image, provided by five well-selected atlas images (mean = 0.6497) was more similar to coverage provided by the entire PA set (mean = 0.6658) than randomly chosen atlas subsets (mean = 0.5977). This was true both of the mean values and the shape of the distributions. Retrospective selection of atlases (mean = 0.6497) provided a very small improvement over prospective atlas selection (mean = 0.6431). All differences were significant (P <1.0E-10). Conclusions Selection of a small representative image set from one dataset can be utilized to develop an atlas set for either retrospective or prospective autosegmentation of a different target dataset. The coverage provided by such a judiciously selected subset has the potential to facilitate propagation of numerous retrospectively defined structures, utilizing additional information available with multimodal imaging in the atlas set, without the need to create large atlas image sets.

AB - Purpose To demonstrate selection of a small representative subset of images from a pool of images comprising a potential atlas (PA) pelvic CT set to be used for autosegmentation of a separate target image set. The aim is to balance the need for the atlas set to represent anatomical diversity with the need to minimize resources required to create a high quality atlas set (such as multiobserver delineation), while retaining access to additional information available for the PA image set. Methods Preprocessing was performed for image standardization, followed by image registration. Clustering was used to select the subset that provided the best coverage of a target dataset as measured by postregistration image intensity similarities. Tests for clustering robustness were performed including repeated clustering runs using different starting seeds and clustering repeatedly using 90% of the target dataset chosen randomly. Comparisons of coverage of a target set (comprising 711 pelvic CT images) were made for atlas sets of five images (chosen from a PA set of 39 pelvic CT and MR images) (a) at random (averaged over 50 random atlas selections), (b) based solely on image similarities within the PA set (representing prospective atlas development), (c) based on similarities within the PA set and between the PA and target dataset (representing retrospective atlas development). Comparisons were also made to coverage provided by the entire PA set of 39 images. Results Exemplar selection was highly robust with exemplar selection results being unaffected by choice of starting seed with very occasional change to one of the exemplar choices when the target set was reduced. Coverage of the target set, as measured by best normalized cross-correlation similarity of target images to any exemplar image, provided by five well-selected atlas images (mean = 0.6497) was more similar to coverage provided by the entire PA set (mean = 0.6658) than randomly chosen atlas subsets (mean = 0.5977). This was true both of the mean values and the shape of the distributions. Retrospective selection of atlases (mean = 0.6497) provided a very small improvement over prospective atlas selection (mean = 0.6431). All differences were significant (P <1.0E-10). Conclusions Selection of a small representative image set from one dataset can be utilized to develop an atlas set for either retrospective or prospective autosegmentation of a different target dataset. The coverage provided by such a judiciously selected subset has the potential to facilitate propagation of numerous retrospectively defined structures, utilizing additional information available with multimodal imaging in the atlas set, without the need to create large atlas image sets.

KW - autosegmentation

KW - clustering

KW - image-atlas

KW - image registration

KW - REGISTRATION

KW - THERAPY

U2 - 10.1002/mp.13494

DO - 10.1002/mp.13494

M3 - Article

VL - 46

SP - 2243

EP - 2250

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 5

ER -

Kennedy A, Dowling J, Greer PB, Holloway L, Jameson MG, Roach D et al. Similarity clustering-based atlas selection for pelvic CT image segmentation. Medical Physics. 2019 May;46(5):2243-2250. https://doi.org/10.1002/mp.13494