Population based ranking of frameless CT-MRI registration methods

Gabor Opposits, Sándor A. Kis, L. Trón, Ervin Berényi, E. Takács, József G. Dobai, L. Bognár, Bernadett Szucs, M. Emri

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Background: Clinical practice often requires simultaneous information obtained by two different imaging modalities. Registration algorithms are commonly used for this purpose. Automated procedures are very helpful in cases when the same kind of registration has to be performed on images of a high number of subjects. Radiotherapists would prefer to use the best automated method to assist therapy planning, however there are not accepted procedures for ranking the different registration algorithms. Purpose: We were interested in developing a method to measure the population level performance of CT-MRI registration algorithms by a parameter of values in the [0,1] interval. Materials and Methods: Pairs of CT and MRI images were collected from 1051 subjects. Results of an automated registration were corrected manually until a radiologist and a neurosurgeon expert both accepted the result as good. This way 1051 registered MRI images were produced by the same pair of experts to be used as gold standards for the evaluation of the performance of other registration algorithms. Pearson correlation coefficient, mutual information, normalized mutual information, Kullback-Leibler divergence, L1 norm and square L2 norm (dis)similarity measures were tested for sensitivity to indicate the extent of (dis)similarity of a pair of individual mismatched images. Results: The square Hellinger distance proved suitable to grade the performance of registration algorithms at population level providing the developers with a valuable tool to rank algorithms. Conclusions: The developed procedure provides an objective method to find the registration algorithm performing the best on the population level out of newly constructed or available preselected ones.

Original languageEnglish
Pages (from-to)353-367
Number of pages15
JournalZeitschrift fur Medizinische Physik
Volume25
Issue number4
DOIs
Publication statusPublished - Dec 1 2015

Fingerprint

Population
Therapeutics

Keywords

  • Computer applications-general
  • CT
  • Medical image registration
  • MRI
  • Segmentation

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology

Cite this

Population based ranking of frameless CT-MRI registration methods. / Opposits, Gabor; Kis, Sándor A.; Trón, L.; Berényi, Ervin; Takács, E.; Dobai, József G.; Bognár, L.; Szucs, Bernadett; Emri, M.

In: Zeitschrift fur Medizinische Physik, Vol. 25, No. 4, 01.12.2015, p. 353-367.

Research output: Contribution to journalArticle

Opposits, Gabor ; Kis, Sándor A. ; Trón, L. ; Berényi, Ervin ; Takács, E. ; Dobai, József G. ; Bognár, L. ; Szucs, Bernadett ; Emri, M. / Population based ranking of frameless CT-MRI registration methods. In: Zeitschrift fur Medizinische Physik. 2015 ; Vol. 25, No. 4. pp. 353-367.
@article{2db4dff34c914e5a887a81e9c66dcfb3,
title = "Population based ranking of frameless CT-MRI registration methods",
abstract = "Background: Clinical practice often requires simultaneous information obtained by two different imaging modalities. Registration algorithms are commonly used for this purpose. Automated procedures are very helpful in cases when the same kind of registration has to be performed on images of a high number of subjects. Radiotherapists would prefer to use the best automated method to assist therapy planning, however there are not accepted procedures for ranking the different registration algorithms. Purpose: We were interested in developing a method to measure the population level performance of CT-MRI registration algorithms by a parameter of values in the [0,1] interval. Materials and Methods: Pairs of CT and MRI images were collected from 1051 subjects. Results of an automated registration were corrected manually until a radiologist and a neurosurgeon expert both accepted the result as good. This way 1051 registered MRI images were produced by the same pair of experts to be used as gold standards for the evaluation of the performance of other registration algorithms. Pearson correlation coefficient, mutual information, normalized mutual information, Kullback-Leibler divergence, L1 norm and square L2 norm (dis)similarity measures were tested for sensitivity to indicate the extent of (dis)similarity of a pair of individual mismatched images. Results: The square Hellinger distance proved suitable to grade the performance of registration algorithms at population level providing the developers with a valuable tool to rank algorithms. Conclusions: The developed procedure provides an objective method to find the registration algorithm performing the best on the population level out of newly constructed or available preselected ones.",
keywords = "Computer applications-general, CT, Medical image registration, MRI, Segmentation",
author = "Gabor Opposits and Kis, {S{\'a}ndor A.} and L. Tr{\'o}n and Ervin Ber{\'e}nyi and E. Tak{\'a}cs and Dobai, {J{\'o}zsef G.} and L. Bogn{\'a}r and Bernadett Szucs and M. Emri",
year = "2015",
month = "12",
day = "1",
doi = "10.1016/j.zemedi.2015.07.001",
language = "English",
volume = "25",
pages = "353--367",
journal = "Zeitschrift fur Medizinische Physik",
issn = "0939-3889",
publisher = "Urban und Fischer Verlag Jena",
number = "4",

}

TY - JOUR

T1 - Population based ranking of frameless CT-MRI registration methods

AU - Opposits, Gabor

AU - Kis, Sándor A.

AU - Trón, L.

AU - Berényi, Ervin

AU - Takács, E.

AU - Dobai, József G.

AU - Bognár, L.

AU - Szucs, Bernadett

AU - Emri, M.

PY - 2015/12/1

Y1 - 2015/12/1

N2 - Background: Clinical practice often requires simultaneous information obtained by two different imaging modalities. Registration algorithms are commonly used for this purpose. Automated procedures are very helpful in cases when the same kind of registration has to be performed on images of a high number of subjects. Radiotherapists would prefer to use the best automated method to assist therapy planning, however there are not accepted procedures for ranking the different registration algorithms. Purpose: We were interested in developing a method to measure the population level performance of CT-MRI registration algorithms by a parameter of values in the [0,1] interval. Materials and Methods: Pairs of CT and MRI images were collected from 1051 subjects. Results of an automated registration were corrected manually until a radiologist and a neurosurgeon expert both accepted the result as good. This way 1051 registered MRI images were produced by the same pair of experts to be used as gold standards for the evaluation of the performance of other registration algorithms. Pearson correlation coefficient, mutual information, normalized mutual information, Kullback-Leibler divergence, L1 norm and square L2 norm (dis)similarity measures were tested for sensitivity to indicate the extent of (dis)similarity of a pair of individual mismatched images. Results: The square Hellinger distance proved suitable to grade the performance of registration algorithms at population level providing the developers with a valuable tool to rank algorithms. Conclusions: The developed procedure provides an objective method to find the registration algorithm performing the best on the population level out of newly constructed or available preselected ones.

AB - Background: Clinical practice often requires simultaneous information obtained by two different imaging modalities. Registration algorithms are commonly used for this purpose. Automated procedures are very helpful in cases when the same kind of registration has to be performed on images of a high number of subjects. Radiotherapists would prefer to use the best automated method to assist therapy planning, however there are not accepted procedures for ranking the different registration algorithms. Purpose: We were interested in developing a method to measure the population level performance of CT-MRI registration algorithms by a parameter of values in the [0,1] interval. Materials and Methods: Pairs of CT and MRI images were collected from 1051 subjects. Results of an automated registration were corrected manually until a radiologist and a neurosurgeon expert both accepted the result as good. This way 1051 registered MRI images were produced by the same pair of experts to be used as gold standards for the evaluation of the performance of other registration algorithms. Pearson correlation coefficient, mutual information, normalized mutual information, Kullback-Leibler divergence, L1 norm and square L2 norm (dis)similarity measures were tested for sensitivity to indicate the extent of (dis)similarity of a pair of individual mismatched images. Results: The square Hellinger distance proved suitable to grade the performance of registration algorithms at population level providing the developers with a valuable tool to rank algorithms. Conclusions: The developed procedure provides an objective method to find the registration algorithm performing the best on the population level out of newly constructed or available preselected ones.

KW - Computer applications-general

KW - CT

KW - Medical image registration

KW - MRI

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=84952873337&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84952873337&partnerID=8YFLogxK

U2 - 10.1016/j.zemedi.2015.07.001

DO - 10.1016/j.zemedi.2015.07.001

M3 - Article

C2 - 26279292

AN - SCOPUS:84952873337

VL - 25

SP - 353

EP - 367

JO - Zeitschrift fur Medizinische Physik

JF - Zeitschrift fur Medizinische Physik

SN - 0939-3889

IS - 4

ER -