Evaluation of deformable registration of patient lung 4DCT with subanatomical region segmentations

Abstract

Deformable registration is needed for a variety of tasks in establishing the voxel correspondence between respiratory phases. Most registration algorithms assume or imply that the deformation field is smooth and continuous everywhere. However, the lungs are contained within closed invaginated sacs called pleurae and are allowed to slide almost independently along the chest wall. This sliding motion is characterized by a discontinuous vector field, which cannot be generated using standard deformable registration methods. The authors have developed a registration method that can create discontinuous vector fields at the boundaries of anatomical subregions. Registration is performed independently on each subregion, with a boundary-matching penalty used to prevent gaps. This method was implemented and tested using both the B-spline and Demons registration algorithms in the Insight Segmentation and Registration Toolkit. The authors have validated this method on four patient 4DCT data sets for registration of the end-inhalation and end-exhalation volumes. Multiple experts identified homologous points in the lungs and along the ribs in the two respiratory phases. Statistical analyses of the mismatch of the homologous points before and after registration demonstrated improved overall accuracy for both algorithms.

Assessment of a commercially available automatic deformable registration system

“…( 3 , 4 , 5 , 6 , 7 ) Registration of thoracic CT images acquired during different respiratory phases allows for modeling of respiratory motion and improved delineation of target margins. For simplicity we register longitudinal thoracic CT ($512\times 512\times 43$ (median) voxels; $0.8\times 0.8\times 7.5\hspace{0.17em}{\text{mm}}^3$) studies imaged on a diagnostic quad‐slice Mx8000 scanner (Philips Medical Systems), which has a measured resolution of approximately 0.7 mm.…”

“…It has been applied to problems such as autosegmentation, ( 1 , 2 ) four‐dimensional (4D) treatment optimization, ( 3 , 4 , 5 , 6 , 7 ) dose accumulation ( 1 , 8 ) and tumor growth/regression analysis. ( 9 ) Varied objectives in these and other potential applications, compounded by the fact that often the lack of a gold standard makes true assessment impossible, ( 10 ) make validation of deformable image registration algorithms very difficult.…”

“…Measuring distances between manually identified points such as vascular and bronchial bifurcations is a common approach to validating thoracic CT deformable registration. ( 3 , 4 , 5 , 6 , 7 ) Brock et al ( 28 ) also identified landmarks on thoracic and abdominal MR images when validating a finite element model (FEM)‐based deformable registration system. Various measures of volume overlap evaluated using radiation oncologist delineated prostate contours have been used for validation of deformable registration of ${\mathrm{normalT}}_2$‐weighted pelvic images with and without inflated endorectal coils.…”

A measure to evaluate deformable registration fields in clinical settings

“…One approach to evaluate a deformable registration technology is to measure accuracy by using markers or clearly visible anatomical locations similar to concepts developed for rigid registration 8 , 12 , 18 , 26 , 30 . Such methods are tedious in clinical practice and incomplete when applied to deformable registration because the methods do not validate how the deformation field models voxels movement inside structures of uniform contrast (31) .…”

Statistical Motion Mask and Sliding Registration

“…Several registration methods for handling sliding motion have been proposed in the literature [1,5,7,10,13,14]. Most of these methods require a segmentation of the sliding regions either in the target image or in both target and source image.…”