Encrypted login | home

Program Information

Surface-Based Deformable Image Registration for Pleural Photodynamic Therapy

no image available
B LIU

B LIU*, T Zhu , Univ Pennsylvania, Philadelphia, PA

Presentations

SU-E-J-232 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose:Pleural photodynamic therapy (PDT) is used in conjunction with lung-sparing surgical treatment for mesothelioma. An infrared (IR) camera based tracking system (NDI) has been used to improve the uniformity of delivered light as well as to contour the chest cavity. Deformable image registration of this contour with the contours obtained using other modalities such as CT during treatment planning can be useful to determine light dose to critical internal organs during PDT treatment.

Methods:CT images were first imported into Matlab so that structure contour of lung could be obtained via the processes of noise filtrations, segmentation using k-means approach, and contouring using edge detection. The obtained two-dimensional lung contour was then rigidly registered with the contour obtained using the NDI system during the surgery before performing the moving mesh (ALE) based deformable registration in COMSOL Multiphysics. The three-dimensional registration was then obtained via the reconstruction of a series of 2-D results.

Results:The contour of lung/chest cavity can be obtained from both CT and NDI system. A series of 2-dimensional deformable image registration are successfully performed for a range of contour cuts (z, along the patient axis) from -5 to 5 cm. Good agreements among contours are obtained for most slices. The reconstructed 3-D surface contours consequently has a good match between the NDI and deformed CT contour.

Conclusion:Surface contours of lung/chest cavity of the same patient obtained from CT scans and NDI tracking system are obtained and pre-processed for the deformable image registration using moving mesh. The preliminary results show good registration results.

Funding Support, Disclosures, and Conflict of Interest: Funding Support: NIH R01-CA-154562


Contact Email: