[home] [Personal Program] [Help]

---

Session: Neuromuscular – upper extremities
Session starts: Friday 25 January, 10:30
Presentation starts: 11:45
Room: Lecture room 535

---

Ruurd Kuiper (UMC Utrecht)
Marijn van Stralen (UMC Utrecht)
Frank Zijlstra (UMC Utrecht)
Harrie Weinans (UMC Utrecht)
Kasper Roth (Erasmus MC)
Joost Colaris (Erasmus MC)
Ralph Sakkers (UMC Utrecht)
Peter Seevinck (UMC Utrecht)

Abstract:
Purpose: The nature of the radio-ulnar joint introduces large nonlinear deformation fields caused by radio-ulnar rotation and elbow extension-flexion motion. The purpose of this work is to register CT to MR images of the forearm region, subjected to these large interscan deformations. We propose a stepwise approach, by first using a rigid body registration and subsequently applying an interpolation step for the soft tissue registration. The registered images can then be used to assist orthopaedic surgeons in pre-operative planning for radio-ulnar osteotomies or to train machine learning techniques that require paired data, such as methods performing automatic registration or to generate synthetic CT contrast images from MR images. Methods: MRI and CT scans were obtained in 15 patients suffering from both-bone forearm fractures. Dixon water-fat reconstruction was used to reconstruct water, fat and in-phase images. A two-step workflow was designed consisting of: 1) deformable BSpline registrations, initialized using rigid body transformations for bone tissue, and 2) a dual quaternion based interpolation scheme for soft tissue. Rigid body transformations were calculated using an Iterative Closest Point (ICP) algorithm applied to segmentations of the humerus, radius and ulna. Results of this workflow were compared to a traditional workflow consisting of rigid and deformable BSpline registration peformed with the widely used software package elastix. The registrations were evaluated quantitatively by segmenting water, fat and bone tissue on both CT and Dixon-reconstructed MR images in order to calculate Dice similarity coefficient. Results: Preliminary results using the proposed workflow showed a Dice coefficient of 0.79, 0.68 and 0.79 for the water, fat and bone tissue respectively, averaging 0.75 over all tissue. For comparison, Dice coefficients using the traditional workflow were 0.74, 0.68 and 0.42, averaging 0.62. Conclusion: The preliminary results showed an increase in Dice coefficient when using the proposed method. Due to the lack of a gold standard for the segmentation of the CT and MR images the results could only be interpreted relative to each other. Qualitative inspection of the results of the proposed workflow indicate regions with considerable registration errors at locations that are distant from the bone. Registration errors also occur close to the edge of the image, caused by a difference in field of view between the CT and MR images. However, tissue close to the bone is generally well registered.