Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH)

Phase I: Segmentation

Philipp Berg, Samuel Voß, Sylvia Saalfeld, Gábor Janiga, Aslak W. Bergersen, Kristian Valen-Sendstad, Jan Bruening, Leonid Goubergrits, Andreas Spuler, Nicole M. Cancelliere, David A. Steinman, Vitor M. Pereira, Tin Lok Chiu, Anderson Chun On Tsang, Bong Jae Chung, Juan R. Cebral, Salvatore Cito, Jordi Pallarès, Gabriele Copelli, Benjamin Csippa & 31 others György Paál, Soichiro Fujimura, Hiroyuki Takao, Simona Hodis, Georg Hille, Christof Karmonik, Saba Elias, Kerstin Kellermann, Muhammad Owais Khan, Alison L. Marsden, Hernán G. Morales, Senol Piskin, Ender A. Finol, Mariya Pravdivtseva, Hamidreza Rajabzadeh-Oghaz, Nikhil Paliwal, Hui Meng, Santhosh Seshadhri, Matthew Howard, Masaaki Shojima, Shin ichiro Sugiyama, Kuniyasu Niizuma, Sergey Sindeev, Sergey Frolov, Thomas Wagner, Alexander Brawanski, Yi Qian, Yu An Wu, Kent D. Carlson, Dan Dragomir-Daescu, Oliver Beuing

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Purpose: Advanced morphology analysis and image-based hemodynamic simulations are increasingly used to assess the rupture risk of intracranial aneurysms (IAs). However, the accuracy of those results strongly depends on the quality of the vessel wall segmentation. Methods: To evaluate state-of-the-art segmentation approaches, the Multiple Aneurysms AnaTomy CHallenge (MATCH) was announced. Participants carried out segmentation in three anonymized 3D DSA datasets (left and right anterior, posterior circulation) of a patient harboring five IAs. Qualitative and quantitative inter-group comparisons were carried out with respect to aneurysm volumes and ostia. Further, over- and undersegmentation were evaluated based on highly resolved 2D images. Finally, clinically relevant morphological parameters were calculated. Results: Based on the contributions of 26 participating groups, the findings reveal that no consensus regarding segmentation software or underlying algorithms exists. Qualitative similarity of the aneurysm representations was obtained. However, inter-group differences occurred regarding the luminal surface quality, number of vessel branches considered, aneurysm volumes (up to 20%) and ostium surface areas (up to 30%). Further, a systematic oversegmentation of the 3D surfaces was observed with a difference of approximately 10% to the highly resolved 2D reference image. Particularly, the neck of the ruptured aneurysm was overrepresented by all groups except for one. Finally, morphology parameters (e.g., undulation and non-sphericity) varied up to 25%. Conclusions: MATCH provides an overview of segmentation methodologies for IAs and highlights the variability of surface reconstruction. Further, the study emphasizes the need for careful processing of initial segmentation results for a realistic assessment of clinically relevant morphological parameters.

Original languageEnglish
Pages (from-to)565-581
Number of pages17
JournalCardiovascular Engineering and Technology
Volume9
Issue number4
DOIs
StatePublished - 15 Dec 2018

Fingerprint

Aneurysm
Anatomy
Intracranial Aneurysm
Surface reconstruction
Hemodynamics
Surface properties
Processing
Ruptured Aneurysm
Rupture
Software

Keywords

  • Challenge
  • Intracranial aneurysm
  • Morphology
  • Segmentation

Cite this

Berg, P., Voß, S., Saalfeld, S., Janiga, G., Bergersen, A. W., Valen-Sendstad, K., ... Beuing, O. (2018). Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH): Phase I: Segmentation. Cardiovascular Engineering and Technology, 9(4), 565-581. https://doi.org/10.1007/s13239-018-00376-0
Berg, Philipp ; Voß, Samuel ; Saalfeld, Sylvia ; Janiga, Gábor ; Bergersen, Aslak W. ; Valen-Sendstad, Kristian ; Bruening, Jan ; Goubergrits, Leonid ; Spuler, Andreas ; Cancelliere, Nicole M. ; Steinman, David A. ; Pereira, Vitor M. ; Chiu, Tin Lok ; Tsang, Anderson Chun On ; Chung, Bong Jae ; Cebral, Juan R. ; Cito, Salvatore ; Pallarès, Jordi ; Copelli, Gabriele ; Csippa, Benjamin ; Paál, György ; Fujimura, Soichiro ; Takao, Hiroyuki ; Hodis, Simona ; Hille, Georg ; Karmonik, Christof ; Elias, Saba ; Kellermann, Kerstin ; Khan, Muhammad Owais ; Marsden, Alison L. ; Morales, Hernán G. ; Piskin, Senol ; Finol, Ender A. ; Pravdivtseva, Mariya ; Rajabzadeh-Oghaz, Hamidreza ; Paliwal, Nikhil ; Meng, Hui ; Seshadhri, Santhosh ; Howard, Matthew ; Shojima, Masaaki ; Sugiyama, Shin ichiro ; Niizuma, Kuniyasu ; Sindeev, Sergey ; Frolov, Sergey ; Wagner, Thomas ; Brawanski, Alexander ; Qian, Yi ; Wu, Yu An ; Carlson, Kent D. ; Dragomir-Daescu, Dan ; Beuing, Oliver. / Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH) : Phase I: Segmentation. In: Cardiovascular Engineering and Technology. 2018 ; Vol. 9, No. 4. pp. 565-581.
@article{062c9943a8a649218f674e721a908de0,
title = "Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH): Phase I: Segmentation",
abstract = "Purpose: Advanced morphology analysis and image-based hemodynamic simulations are increasingly used to assess the rupture risk of intracranial aneurysms (IAs). However, the accuracy of those results strongly depends on the quality of the vessel wall segmentation. Methods: To evaluate state-of-the-art segmentation approaches, the Multiple Aneurysms AnaTomy CHallenge (MATCH) was announced. Participants carried out segmentation in three anonymized 3D DSA datasets (left and right anterior, posterior circulation) of a patient harboring five IAs. Qualitative and quantitative inter-group comparisons were carried out with respect to aneurysm volumes and ostia. Further, over- and undersegmentation were evaluated based on highly resolved 2D images. Finally, clinically relevant morphological parameters were calculated. Results: Based on the contributions of 26 participating groups, the findings reveal that no consensus regarding segmentation software or underlying algorithms exists. Qualitative similarity of the aneurysm representations was obtained. However, inter-group differences occurred regarding the luminal surface quality, number of vessel branches considered, aneurysm volumes (up to 20{\%}) and ostium surface areas (up to 30{\%}). Further, a systematic oversegmentation of the 3D surfaces was observed with a difference of approximately 10{\%} to the highly resolved 2D reference image. Particularly, the neck of the ruptured aneurysm was overrepresented by all groups except for one. Finally, morphology parameters (e.g., undulation and non-sphericity) varied up to 25{\%}. Conclusions: MATCH provides an overview of segmentation methodologies for IAs and highlights the variability of surface reconstruction. Further, the study emphasizes the need for careful processing of initial segmentation results for a realistic assessment of clinically relevant morphological parameters.",
keywords = "Challenge, Intracranial aneurysm, Morphology, Segmentation",
author = "Philipp Berg and Samuel Vo{\ss} and Sylvia Saalfeld and G{\'a}bor Janiga and Bergersen, {Aslak W.} and Kristian Valen-Sendstad and Jan Bruening and Leonid Goubergrits and Andreas Spuler and Cancelliere, {Nicole M.} and Steinman, {David A.} and Pereira, {Vitor M.} and Chiu, {Tin Lok} and Tsang, {Anderson Chun On} and Chung, {Bong Jae} and Cebral, {Juan R.} and Salvatore Cito and Jordi Pallar{\`e}s and Gabriele Copelli and Benjamin Csippa and Gy{\"o}rgy Pa{\'a}l and Soichiro Fujimura and Hiroyuki Takao and Simona Hodis and Georg Hille and Christof Karmonik and Saba Elias and Kerstin Kellermann and Khan, {Muhammad Owais} and Marsden, {Alison L.} and Morales, {Hern{\'a}n G.} and Senol Piskin and Finol, {Ender A.} and Mariya Pravdivtseva and Hamidreza Rajabzadeh-Oghaz and Nikhil Paliwal and Hui Meng and Santhosh Seshadhri and Matthew Howard and Masaaki Shojima and Sugiyama, {Shin ichiro} and Kuniyasu Niizuma and Sergey Sindeev and Sergey Frolov and Thomas Wagner and Alexander Brawanski and Yi Qian and Wu, {Yu An} and Carlson, {Kent D.} and Dan Dragomir-Daescu and Oliver Beuing",
year = "2018",
month = "12",
day = "15",
doi = "10.1007/s13239-018-00376-0",
language = "English",
volume = "9",
pages = "565--581",
journal = "Cardiovascular Engineering and Technology",
issn = "1869-408X",
number = "4",

}

Berg, P, Voß, S, Saalfeld, S, Janiga, G, Bergersen, AW, Valen-Sendstad, K, Bruening, J, Goubergrits, L, Spuler, A, Cancelliere, NM, Steinman, DA, Pereira, VM, Chiu, TL, Tsang, ACO, Chung, BJ, Cebral, JR, Cito, S, Pallarès, J, Copelli, G, Csippa, B, Paál, G, Fujimura, S, Takao, H, Hodis, S, Hille, G, Karmonik, C, Elias, S, Kellermann, K, Khan, MO, Marsden, AL, Morales, HG, Piskin, S, Finol, EA, Pravdivtseva, M, Rajabzadeh-Oghaz, H, Paliwal, N, Meng, H, Seshadhri, S, Howard, M, Shojima, M, Sugiyama, SI, Niizuma, K, Sindeev, S, Frolov, S, Wagner, T, Brawanski, A, Qian, Y, Wu, YA, Carlson, KD, Dragomir-Daescu, D & Beuing, O 2018, 'Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH): Phase I: Segmentation', Cardiovascular Engineering and Technology, vol. 9, no. 4, pp. 565-581. https://doi.org/10.1007/s13239-018-00376-0

Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH) : Phase I: Segmentation. / Berg, Philipp; Voß, Samuel; Saalfeld, Sylvia; Janiga, Gábor; Bergersen, Aslak W.; Valen-Sendstad, Kristian; Bruening, Jan; Goubergrits, Leonid; Spuler, Andreas; Cancelliere, Nicole M.; Steinman, David A.; Pereira, Vitor M.; Chiu, Tin Lok; Tsang, Anderson Chun On; Chung, Bong Jae; Cebral, Juan R.; Cito, Salvatore; Pallarès, Jordi; Copelli, Gabriele; Csippa, Benjamin; Paál, György; Fujimura, Soichiro; Takao, Hiroyuki; Hodis, Simona; Hille, Georg; Karmonik, Christof; Elias, Saba; Kellermann, Kerstin; Khan, Muhammad Owais; Marsden, Alison L.; Morales, Hernán G.; Piskin, Senol; Finol, Ender A.; Pravdivtseva, Mariya; Rajabzadeh-Oghaz, Hamidreza; Paliwal, Nikhil; Meng, Hui; Seshadhri, Santhosh; Howard, Matthew; Shojima, Masaaki; Sugiyama, Shin ichiro; Niizuma, Kuniyasu; Sindeev, Sergey; Frolov, Sergey; Wagner, Thomas; Brawanski, Alexander; Qian, Yi; Wu, Yu An; Carlson, Kent D.; Dragomir-Daescu, Dan; Beuing, Oliver.

In: Cardiovascular Engineering and Technology, Vol. 9, No. 4, 15.12.2018, p. 565-581.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH)

T2 - Phase I: Segmentation

AU - Berg, Philipp

AU - Voß, Samuel

AU - Saalfeld, Sylvia

AU - Janiga, Gábor

AU - Bergersen, Aslak W.

AU - Valen-Sendstad, Kristian

AU - Bruening, Jan

AU - Goubergrits, Leonid

AU - Spuler, Andreas

AU - Cancelliere, Nicole M.

AU - Steinman, David A.

AU - Pereira, Vitor M.

AU - Chiu, Tin Lok

AU - Tsang, Anderson Chun On

AU - Chung, Bong Jae

AU - Cebral, Juan R.

AU - Cito, Salvatore

AU - Pallarès, Jordi

AU - Copelli, Gabriele

AU - Csippa, Benjamin

AU - Paál, György

AU - Fujimura, Soichiro

AU - Takao, Hiroyuki

AU - Hodis, Simona

AU - Hille, Georg

AU - Karmonik, Christof

AU - Elias, Saba

AU - Kellermann, Kerstin

AU - Khan, Muhammad Owais

AU - Marsden, Alison L.

AU - Morales, Hernán G.

AU - Piskin, Senol

AU - Finol, Ender A.

AU - Pravdivtseva, Mariya

AU - Rajabzadeh-Oghaz, Hamidreza

AU - Paliwal, Nikhil

AU - Meng, Hui

AU - Seshadhri, Santhosh

AU - Howard, Matthew

AU - Shojima, Masaaki

AU - Sugiyama, Shin ichiro

AU - Niizuma, Kuniyasu

AU - Sindeev, Sergey

AU - Frolov, Sergey

AU - Wagner, Thomas

AU - Brawanski, Alexander

AU - Qian, Yi

AU - Wu, Yu An

AU - Carlson, Kent D.

AU - Dragomir-Daescu, Dan

AU - Beuing, Oliver

PY - 2018/12/15

Y1 - 2018/12/15

N2 - Purpose: Advanced morphology analysis and image-based hemodynamic simulations are increasingly used to assess the rupture risk of intracranial aneurysms (IAs). However, the accuracy of those results strongly depends on the quality of the vessel wall segmentation. Methods: To evaluate state-of-the-art segmentation approaches, the Multiple Aneurysms AnaTomy CHallenge (MATCH) was announced. Participants carried out segmentation in three anonymized 3D DSA datasets (left and right anterior, posterior circulation) of a patient harboring five IAs. Qualitative and quantitative inter-group comparisons were carried out with respect to aneurysm volumes and ostia. Further, over- and undersegmentation were evaluated based on highly resolved 2D images. Finally, clinically relevant morphological parameters were calculated. Results: Based on the contributions of 26 participating groups, the findings reveal that no consensus regarding segmentation software or underlying algorithms exists. Qualitative similarity of the aneurysm representations was obtained. However, inter-group differences occurred regarding the luminal surface quality, number of vessel branches considered, aneurysm volumes (up to 20%) and ostium surface areas (up to 30%). Further, a systematic oversegmentation of the 3D surfaces was observed with a difference of approximately 10% to the highly resolved 2D reference image. Particularly, the neck of the ruptured aneurysm was overrepresented by all groups except for one. Finally, morphology parameters (e.g., undulation and non-sphericity) varied up to 25%. Conclusions: MATCH provides an overview of segmentation methodologies for IAs and highlights the variability of surface reconstruction. Further, the study emphasizes the need for careful processing of initial segmentation results for a realistic assessment of clinically relevant morphological parameters.

AB - Purpose: Advanced morphology analysis and image-based hemodynamic simulations are increasingly used to assess the rupture risk of intracranial aneurysms (IAs). However, the accuracy of those results strongly depends on the quality of the vessel wall segmentation. Methods: To evaluate state-of-the-art segmentation approaches, the Multiple Aneurysms AnaTomy CHallenge (MATCH) was announced. Participants carried out segmentation in three anonymized 3D DSA datasets (left and right anterior, posterior circulation) of a patient harboring five IAs. Qualitative and quantitative inter-group comparisons were carried out with respect to aneurysm volumes and ostia. Further, over- and undersegmentation were evaluated based on highly resolved 2D images. Finally, clinically relevant morphological parameters were calculated. Results: Based on the contributions of 26 participating groups, the findings reveal that no consensus regarding segmentation software or underlying algorithms exists. Qualitative similarity of the aneurysm representations was obtained. However, inter-group differences occurred regarding the luminal surface quality, number of vessel branches considered, aneurysm volumes (up to 20%) and ostium surface areas (up to 30%). Further, a systematic oversegmentation of the 3D surfaces was observed with a difference of approximately 10% to the highly resolved 2D reference image. Particularly, the neck of the ruptured aneurysm was overrepresented by all groups except for one. Finally, morphology parameters (e.g., undulation and non-sphericity) varied up to 25%. Conclusions: MATCH provides an overview of segmentation methodologies for IAs and highlights the variability of surface reconstruction. Further, the study emphasizes the need for careful processing of initial segmentation results for a realistic assessment of clinically relevant morphological parameters.

KW - Challenge

KW - Intracranial aneurysm

KW - Morphology

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85057759305&partnerID=8YFLogxK

U2 - 10.1007/s13239-018-00376-0

DO - 10.1007/s13239-018-00376-0

M3 - Article

VL - 9

SP - 565

EP - 581

JO - Cardiovascular Engineering and Technology

JF - Cardiovascular Engineering and Technology

SN - 1869-408X

IS - 4

ER -

Berg P, Voß S, Saalfeld S, Janiga G, Bergersen AW, Valen-Sendstad K et al. Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH): Phase I: Segmentation. Cardiovascular Engineering and Technology. 2018 Dec 15;9(4):565-581. https://doi.org/10.1007/s13239-018-00376-0