In-room breathing motion estimation from limited projection views using a sliding deformation model - CNRS - Centre national de la recherche scientifique Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

In-room breathing motion estimation from limited projection views using a sliding deformation model

Résumé

Purpose: To estimate in-room breathing motion from a limited number of 2D cone-beam projection images by deformably registering them to a phase of the planning 4DCT. Methods: Breathing motion was modelled using a piecewise continuous B-spline representation, allowing to preserve the sliding along the thoracic wall while limiting the degrees of freedom. The deformed target 3D image was subsequently used to generate Digitally Reconstructed Radiographs. Their normalised correlation coefficient with the measured projection images was computed in the 2D projection space but its derivatives were backprojected into the 3D space, avoiding the projection of the transform jacobian matrix which is computationally intractable. Results: The method was qualitatively evaluated on a lung cancer patient treated in the Léon Bérard Cancer Center. The deformed planning CT visually matched well the reconstructed cone-beam CT. Quantification was done on 2 patients of the DIR-labs database with different breathing amplitudes. Forty cone-beam projection images were simulated using the end-exhale phase of the planning 4DCT and the detector positions of a real middle field of view cone-beam acquisition (off-centered). The end-inhale phase was deformed to match these simulated projections. The Target Registration Error decreased from 3.89 mm to 1.14 mm for the first patient and from 15 mm to 4.21 mm for the second. For the first patient, the 3D/3D registration of the reconstructed cone-beam CT obtained with the same set of simulated projections was comparatively worse (1.46 mm), due to view aliasing artefacts. Conclusions: 2D/3D deformable registration seems to improve the quality of in-room breathing motion estimation from a limited number of views

Dates et versions

hal-00847368 , version 1 (23-07-2013)

Identifiants

Citer

V. Delmon, J. Vandemeulebroucke, R. Pinho, M. Vila Oliva, D. Sarrut, et al.. In-room breathing motion estimation from limited projection views using a sliding deformation model. International Conference on the Use of Computers in Radiation Therapy (ICCR) 2013, May 2013, Melbourne, Australia. pp.012026, ⟨10.1088/1742-6596/489/1/012026⟩. ⟨hal-00847368⟩
83 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More