Comparison of different segmentation approaches without using gold standard. Application to the estimation of the left ventricle ejection fraction from cardiac cine MRI sequences.

Jessica Lebenberg (1) , Irène Buvat (2) , Mireille Garreau (3) , Christopher Casta (4) , Constantin Constantinidès (1) , Jean Cousty (5) , Alexandre Cochet (6) , Stéphanie Jehan-Besson (7) , Christophe Tilmant (8) , Muriel Lefort (1) , Elodie Roullot (9) , Laurent Najman (5) , Laurent Sarry (10) , Patrick Clarysse (4) , Alain de Cesare (1) , Alain Lalande (6) , Frédérique Frouin (1)

1 LIF - Laboratoire d'Imagerie Fonctionnelle
2 IMNC (UMR_8165) - Imagerie et Modélisation en Neurobiologie et Cancérologie
3 LTSI - Laboratoire Traitement du Signal et de l'Image
4 MOTIVATE - Imagerie et modélisation Vasculaires, Thoraciques et Cérébrales
5 LIGM - Laboratoire d'Informatique Gaspard-Monge
6 Le2i - Laboratoire Electronique, Informatique et Image [UMR6306]
7 Equipe Image - Laboratoire GREYC - UMR6072
8 LASMEA - Laboratoire des sciences et matériaux pour l'électronique et d'automatique
9 PRIAM
10 ISIT - Image Science for Interventional Techniques

A statistical method is proposed to compare several estimates of a relevant clinical parameter when no gold standard is available. The method is illustrated by considering the left ventricle ejection fraction derived from cardiac magnetic resonance images and computed using seven approaches with different degrees of automation. The proposed method did not use any a priori regarding with the reliability of each method and its degree of automation. The results showed that the most accurate estimates of the ejection fraction were obtained using manual segmentations, followed by the semiautomatic methods, while the methods with the least user input yielded the least accurate ejection fraction estimates. These results were consistent with the expected performance of the estimation methods, suggesting that the proposed statistical approach might be helpful to assess the performance of estimation methods on clinical data for which no gold standard is available.

Domains: Bioengineering; Signal and Image processing; Signal and Image Processing

Jessica Lebenberg, Irène Buvat, Mireille Garreau, Christopher Casta, Constantin Constantinidès, et al.. Comparison of different segmentation approaches without using gold standard. Application to the estimation of the left ventricle ejection fraction from cardiac cine MRI sequences.. Annual International Conference of the IEEE Engineering in Medicine and Biology Society., Aug 2011, Boston, United States. pp.2663-6, ⟨10.1109/IEMBS.2011.6090732⟩. ⟨inserm-00773261⟩